At a certain point, your resume should speak for itself. The fact that experienced engineers with impressive resumes are put through these types of interviews is insulting and frustrating to the interviewees.
Succeeding at these whiteboard questions requires weeks of preparation. You need to practice, practice, practice. After enough practice, you are pretty likely to pass. So ultimately, it is more of a test of "how much do you want to work here." If you care enough, you can pass. That said, these types of interviews do favor fresh grads over experienced programmers (you have algorithms and data structures fresh in your mind), which means that they are flawed in IMHO.
These types of interviews work really well for the "I got 1600 on my SATs and went to {insert high profile school here} crowd" There are books out there just to prep you for Google interviews I see these as very similar to SAT test prep books. I'm not so sure Google is really interested in hiring the best engineers but rather a specific type of engineer.
Think of the problem from Google's perspective though. At some point, you have tens of thousands of candidates and you need a system to quantify how good they are. Further, it's reasonable to have false negatives (people you don't hire that should have been hired) but really bad to have false positives (people that you hire that you should not have). Together, these boil down into the de facto whiteboarding interview process.
This has got to be the biggest hiring fallacy I've ever heard. "It's better to reject a good candidate than hire a bad candidate."
That's completely false, and anyone who says that is completely ignorant of Bayesian logic.
Here are some simple numbers.
Suppose that a "good" candidate is a 1-in-100 find. Suppose that a "bad" candidate has a 1% chance of tricking you into hiring them anyway.
Every time you pass on a "good" candidate, that is greater opportunity for a "bad" candidate to trick you into hiring them!
Counterintuitively, if you pass on too many of the "good" candidates, in your overzealousness to reject bad candidates, you're actually INCREASING THE ODDS OF A BAD HIRE.
This is management porn. "It's better to reject a good candidate than hire a bad candidate." It makes the manager feel good. "Wow! That candidate seemed smart, but I rejected him anyway! I'm such a great leader! I make the tough decisions!"
tl;dr summary
Because "good" candidates are rare, every time you pass on a good candidate that increases your odds of making a bad hire! This is simple Bayesian reasoning!
Those are simple numbers, but they don't get at the real issue. People say that because 1 bad hire can have negative effects on the whole team, causing others to get less done and leave. Missing a good hire doesn't poison your team.
But 1 bad hire is easily correctable - you fire the person. You won't know that you missed on the good hires.
I personally have worked with several awesome people that have interviewed with Google, and none of them got hired. They all said the interview process was flat out insulting.
Interestingly enough, most of them ended up at Facebook.
>>But 1 bad hire is easily correctable - you fire the person.
Not sure if you have ever been a manager, because firing someone is NEVER easy. It hurts everyone emotionally. The person getting fired feels awful. The person doing the firing feels awful (unless they are a real sociopath). And team morale tends to take a big hit.
Here is my stance: if you hire someone who isn't a good fit, unless they actively deceived you, it is your god damn job to find a way to make it work. As a principle you should treat people with respect and dignity, and not as easily disposable and replaceable cogs.
> it is your god damn job to find a way to make it work.
No, it isn't. This is why probation periods exist. If you hired someone and they aren't working out, then at the end of the probation period you don't keep them on.
Trying to force someone who doesn't fit the team dynamic to stay is going to hurt your org. It doesn't make you a "good" manager to say to everyone "I know this sucks but deal with it because letting them go would mean I was wrong."
> As a principle you should treat people with respect and dignity, and not as easily disposable and replaceable cogs.
Which is ironic because so many job descriptions I see today are basically written as "we want to hire someone with exactly these skills, who requires zero training, and can become an expert in our systems in their first week."
No one is that way, unless your system consists of pressing one button all day, but then the job description would probably require that the person have intricate knowledge of the button and is friends with the engineer who designed it so they know what to do if the button suddenly stops working.
If companies actually treated people with respect and dignity, I'd get the training I need to become a better employee, instead of going to management and begging them for any training every 6 months like the industry is now.
> Trying to force someone who doesn't fit the team dynamic to stay is going to hurt your org.
Doesn't this rather presume that your "team dynamic" is actually good to start with?
Let's say you had a team of not very good engineers. Then you hired quite a good engineer who looked at all the terrible practices (e.g. no version control, shitty or nonexistent testing, poor build processes) and said to themselves "look, I need to fix this shit or I'm leaving, and I've got ten better offers in my inbox".
They might not be a very good team fit, but that's because the team is filled with idiots who haven't figured out how to use version control or whatever.
So, you know, the best thing to do is to get rid of them for not being a culture fit or for not being good for the team dynamic...?
> They might not be a very good team fit, but that's because the team is filled with idiots who haven't figured out how to use version control or whatever.
Yes, if you have a team like this and you hire someone who has a higher standard, there is going to be some friction between them and the existing team members. Letting that person go isn't your concern, because as you said yourself:
> and said to themselves "look, I need to fix this shit or I'm leaving, and I've got ten better offers in my inbox".
I've been in that situation before. I was hired to a company and when I got there I found out that every single day they were fighting fires because of stupid decisions management made with little foresight into how it would affect the team. Funny enough none of this was mentioned during the interview, although it was a definite red flag that they had high churn for this particular position.
I stayed there for my probation period trying to fix things so the team would fire fight every day and change management's mentality, but it wasn't happening, so I gave notice and went somewhere better.
> So, you know, the best thing to do is to get rid of them for not being a culture fit or for not being good for the team dynamic...?
No, you took me too literally. Obviously if you've got someone who has friction with the team, but they're a hard working individual who is trying to make your team better and more efficient, you should try to work through those stressful periods because in the long run it will be better for your team's health and the company's health. If some of your low performers leave during this period, that's okay, they were only going to hurt you in the long run. That being said, don't burn bridges with your existing employees. Try to find a happy middle ground that results in a better work environment for everyone.
I have been a manager, and I have had to fire people.
I have also quit a job because they fired the wrong person.
In my opinion, if you fire someone and the reaction from the whole team isn't "thank god, that guy needed to go", then you made a mistake firing them and should have tried to make it work. In the past when I fired someone, the general reaction was much more along the lines of "what took you so long". In both cases I was the one at the table defending the person while everyone else said they needed to go.
Interestingly, in both cases the reason for firing them was that the person in question was not treating other people with respect. So what is worse, letting one person shit on your whole team and make everyone else feel bad, or getting that person off your team?
"it is your god damn job to find a way to make it work"
That attitude is how bloat happens. There is a middle ground here. Not firing bad employees also hurts many people emotionally as does an underperforming, overbloated company. Employees are as replaceable as employers. To use a sports analogy, it's the big leagues and you may not make the team and often times it is hard to know whether someone can make the team without a trial. I believe these draconian interview practices and contract-to-hire approaches are a direct result of the never-fire-bad-employees attitude.
If Google didn't apply the equivalent of the Hogwarts Hat to placing its employees, I'd agree with you. But in doing so, they remove the ability of the potential Googler to meet their potential teammates and decide whether it's a good fit.
Ergo they own the problem, not so much the Noogler. Also generalists == mediocre code in my experience because there's no passion for such work in a lot of people who are fantastic with other in-demand skills, but I digress.
And when I joined, Google wasn't so much bloating as it was metastasizing. Most of the people I knew who joined around the same time have long since left. In contrast, I've been at the same gig ever since I left Google. The Google experience was the outlier.
Hmm. When I joined Google, I interviewed the managers of both teams I was offered a position on, and a team member on the one I ultimately decided to join.
I'm not sure if this is the norm--I believe it differs for industry hires versus those fresh out of school--but it's certainly not true that, as a rule, you can never meet potential teammates.
Two decades in the valley has taught me that you can't decide anything in a half hour conversation beyond recognizing an absolute bozo incarnate. You need to meet the whole team and preferably have lunch with them and figure out whether it's the right next step. What Google does instead reeks of magical thinking to me.
Because for the most part, Googlers aren't bozos (although a couple of the true believers in the Googleplex blew my mind with how crufty their skills and knowledge had become), but that doesn't rule out one of the many great engineers there promoted to their managerial level of incompetence. And that's what I had - a guy with no people skills whatsoever - and two teams with a total of 20+ engineers to "manage."
What does that even mean? Swap left-right child nodes? Write it linearly (n-th children at 2n and 2n+1) and reverse it as if it was a string? OP later described it as "to min-max the tree, ascending to descending." confusing me further.
>People say that because 1 bad hire can have negative effects on the whole team, causing others to get less done and leave. Missing a good hire doesn't poison your team.
That doesn't disprove what he said. In fact it strengthens it and makes missing a good hire even worse, because, as he said missing good hires -> MORE change of bad hires (that can "poison your team").
In other words, it's not a choice: "I'm better of rejecting a good hire than getting a bad one, because a bad one could poison the team".
By rejecting good ones, who're actually getting MORE bad ones. So it should be corrected to:
"I'm better of NOT rejecting good hires, or else I'll get more bad hires and poison my team".
> Counterintuitively, if you pass on too many of the "good" candidates, in your overzealousness to reject bad candidates, you're actually INCREASING THE ODDS OF A BAD HIRE.
Yes, but what hiring managers are doing isn't passing over candidates that they know are good, they're passing over candidates that they're not very sure about.
> Because "good" candidates are rare, every time you pass on a good candidate that increases your odds of making a bad hire! This is simple Bayesian reasoning!
Similarly, of course this is true, but again hiring managers aren't passing up candidates that they know to be good, they're passing up on candidates that they're not sure about, whom the acknowledge could be good.
Under certain conditions, every time you pass on a candidate you're not sure about, you decrease your chance of making a bad hire. The conditions are that increasing hiring standards must weed out more bad candidates than good candidates.
Let's work with your model. Suppose 1% of candidates coming for an interview are good and 99% are bad.
Hiring strategy A manages to hire all good candidates it interviews, and 1% of the bad candidates it interviews. End result: for every 100 people you interview, you get 1 good candidate and 1 bad candidate.
Hiring strategy B is more conservative and hires 50% of all good candidates it interviews, and 0.1% of the bad candidates it interviews. End result: for every 100 people you interview, you get 0.5 good candidates and 0.1 bad candidates.
The claim is something like, the team produced by strategy B is better than the team produced by strategy A, even though team B has less good candidates than team B.
I agree with you that if hiring strategy B led to hiring 10% of the good candidates and cut the odds of hiring the bad candidate to 0.5%, compared to 100% and 1% for hiring strategy A, then strategy B is worse than strategy A. However
1) This means that by raising your hiring bar (adopting B instead of A), you eliminate more good candidates than bad candidates. You now have to prove this empirical claim.
2) All I wanted to say was that the statements you made like "every time you pass on a good candidate that increases your odds of making a bad hire!" is only true under certain conditions (namely that raising hiring standards eliminates more good candidates than bad candidates), and it's not "simple bayesian reasoning".
The original slogan "It is better to reject a good candidate than hire a bad candidate." specifically says to make that fallacy.
As long as good candidates are rare, and bad candidates have a small-but-nonzero chance of tricking you into hiring them, it's very expensive to pass on a good candidate.
Even if you pass on 50% of the good candidates (no matter how many bad ones trick you), you're doubling the amount of time you're spending interviewing.
The only way to be sure about how many "good" candidates your are missing is to hire some random percentage of the people who fail your interview. Only a large tech company would have the resources to do that, and I'm not sure if it would be ethical, or even get the employer in legal trouble.
NO employer knows how many good candidates they miss out on, which makes this analysis very difficult.
Also, I question the ability of most employers to evaluate employees AFTER they are hired, so any analysis of post-interview performance might just be reinforcing whatever biases are present.
Also, candidates fall into THREE groups: great, mediocre, and toxic. The catch is that the toxic ones are most likely to trick you into thinking they're great. That makes it even more expensive to pass on a good candidate.
> As long as good candidates are rare, and bad candidates have a small-but-nonzero chance of tricking you into hiring them, it's very expensive to pass on a good candidate.
Yes, I agree with you. My point is that even taking this into account, even after taking into account all the other costs you mention into account, it might still be better to pass on candidates you're not sure about.
> The original slogan "It is better to reject a good candidate than hire a bad candidate." specifically says to make that fallacy.
So going by the slogan, even if passing on too many good candidates increases your odds of a bad hire (say from 50% to 80%), it's still better than hiring a bad candidate, since if you hire a bad candidate, the odds that you made a bad hire is 100%!
My interpretation of the slogan seems different from yours, mine is something like: if you hire a good person the value of your business will increase by X, if you hire a bad person the value of your business will decrease by Y, and Y is much greater than X.
A bad hire can easily cause more friction than if nobody was hired at all. Net effect for the business is negative. You're also probably severely overestimating the likelihood of a bad hire going through.
But the cost and other effects of doing so vary significantly depending on where you are. Most of the US may have at-will employment, but much of the rest of the world does not.
But the US is the place where the "It's better to reject a good candidate than hire a bad candidate" mantra comes from! European companies practice it, but they do it because they're forced to, and they don't preach it like a gospel. The US does it without any need to do it, and thinks they're gaining some sort of advantage in the process.
We had a really bad hire about 2 years ago. So bad it caused us to review our entire hiring process to understand how he got in. It took almost a year before he was fired (and he was a contractor so it should have been easier). In that time, he used up untold resources while we tried to find work he could actually do, people helping him "just in case he just needed a helping hand" and so on. Finally after wasting other people's time for 9 months, a manager made the decision to get him out.
A year later I got an email from him asking for a recommendation. Nothing ventured, nothing gained I guess.
Here I see the same logic that caused the TSA. Some rare and bad event happens, and people rush to rebuild the whole system to never let that happen again. Not even considering maybe accepting that rare and bad events happen and dealing with them more efficiently would be better? Maybe rebuild actually doesn't even prevent bad events, just specific type of bad event that already happened and is unlikely to repeat anytime soon? Maybe dealing with imperfection is better than trying to be perfect every time?
I.e., I don't know your particular situation, so I wouldn't dare to assume the reaction was wrong, but too often "we much at all costs make sure it never happens again" overwhelms any attempt at reasonable approach and cost/benefit analysis completely. Maybe getting one bad apple in occasionally is a reasonable price for quickly and efficiently hiring a lot of good people?
Because it's hard to fire someone. Lets discount the emotional cost completely and simply look at this in terms of financial risk to the company.
Employee X is hired. After a month it becomes clear to the whole team that X is a bad hire. But how do you turn a subjectively obvious feeling into an objectively obvious reason to fire. Because if that person decides to sue the company after getting fired that's what you will have to demonstrate.
Just imagine how hard it is to quantify what makes someone good at our job? Code Quality? Team dynamic? These are really fuzzy metrics.
You could document how many times the engineer affected production badly and how much that cost the company. But to be realistic even great engineers do that. Heck I touched production at Google and decreased revenue by millions of dollars and it was expected that this happens. So you'd have to let him affect the company enough times to be obvious He's worse than your other engineers.
You are going to have to have documented enough clear failures to do the job to justify starting the process of firing. If you don't have a performance improvement plan then you might be able to speed up the process a little. However not having a performance improvement plan is both a bad idea and, I would imagine (IANAL etc.), could cause problems later on down the road in a potential suit.
Taken in that light 9 months doesn't seem that long at all.
Idiot: We don't want to make any bad hires! Therefore, we reject a lot of good candidates!
FSK: If you pass on many good candidates, and you have a small chance of hiring any given bad candidate, then each good candidate you reject actually INCREASES your odds of making a bad hire.
Idiot: We made a bad hire once! Never again! Now we reject lots of good candidates to avoid that repeat disaster!
FSK: But, if you want to minimize your bad hire rate, you also have to minimize the number of good candidates you reject.
Idiot: NO! NO! NO! The way you make sure you hire no bad candidates is to be so strict that you reject lots of good ones! That's what everyone told me so it must be right!
In one ear and out the other. Why do people who don't understand statistics get to be managers? If you don't understand this statistics argument I made, you're unqualified to work in any sort of technical area.
It's actually fairly reasonable. The manager doesn't want to avoid hiring bad candidates. The manager wants to avoid being blamed for hiring bad candidates. They want to be able to say "I looked really really hard for reasons why Bob was a bad hire, and didn't find any, so you can't blame me for Bob being a bad hire."
Whether or not this actually reduces the number of bad hires is beside the point. It's a classic principle-agent problem.
You're missing something vital here. They're not rejecting good candidates because they're good candidates, they're rejecting them because they're not sure enough that they're good candidates.
They fear they might be bad candidates. They'd rather reject too many than too few, so they make sure they reject everybody who they're not 100% is a good candidate. That means they may reject some good candidates that aren't easily identified as such, but it doesn't increase their chance of hiring a bad candidate; it decreases it.
Strictly speaking, I think raising the standards also could lower the probability of a bad hire (depends on the model - is bad hire completely random? does it depend on some parameter that is controlled?) together with a probability of a good hire, so I'm not sure it is a robust argument that raising standards always raises chance of a bad hire. I'd be happy to see a more rigorous proof (I know it's complete waste of time but for some people it's fun).
Basing all of your management decisions purely on statistics is a bad strategy. You're assuming that the weighting of good vs. bad candidate is the same, when they're not.
Hiring a good employee is not nearly as impactful as hiring a bad one. If you can have a strategy that filters out 100% of bad employees, but unfortunately also filters out 90% of good employees, this is preferable to filtering only 50% of good employees, but also only filtering 90% of bad employees. You may have more good employees with the latter strategy, but the bad employees can kill the team.
People have an irrational belief in their ability to tell the difference. This is how one gets such flawed thinking in the first place (regarding priors and probability distributions).
Firms are hierarchy-based entities. "Technical Merit" is about a third or fourth order away from what really drives the hiring decision. Its also very imperfect predictor of actual performance.
The issues is that companies use the term all the time. They want people to believe they were hired for their merit, but merit is almost always 'fit' and not technical in nature. The technical hoops are just a CYA for when the 'fit' doesn't work out (mis-judgement) and they need something to point to as to that is not arbitrary in nature to explain how other people were not hired instead.
Not really. The population is large enough that you're treating each potential hire as a random variable. Given the amount of possible good new hires the effect of passing on one actual good hire has negligible effect on the next potential hire.
The theory behind the "raise the bar" argument is that obviously good hires are easy to spot (this is the part most people reasonably take issue with), so people who do not pass that bar of "obviously good" aren't worth taking a risk on. The conclusion is statistically reasonable if you accept the premise (which, again, you probably shouldn't).
> The population is large enough that you're treating each potential hire as a random variable.
This is only true for new grads. The population of engineers with ten years of appropriate corporate experience (that match the technology survivor bias) and at the ~12 locations google can use is tiny.
When I look at my old corporation, only 3 people (of 500+ I would be aware of) were hired by google. One of them is genuinely great, but I doubt he passed the current process as a blind hire. The other two I wouldn't want to work with, but I bet they passed this process precisely because the right kind of focus. (I.e. everything people around them didn't know was their priority and since we normally focus on what's important to completing our project, none of that was so important..)
I was very interested in google when I was under the impression that (without experiencing menlo park on a daily basis,) I could either walk away with over $500k (gross) in 2- years, enjoy working there for 5 years, and/or work with the best >experienced< engineers in our field. When glassdoor and my own linkedin network revealed my best perspective into their setup, blind applying to them is now just a good filter to test my new age employability skills. So, now they've built themselves a new problem of true positive false positives who are just along for the ride.
The hiring market is already NP Hard for both sides when you are genuinely trying to get a good enough fit for yourself rather than the best deal on the market. If you try to misuse the other side they will see less ethical constraints in using the resources you'll have to put in.
How many developers genuinely want a job they can't do that they will have to be awkwardly fired from in six months and explain at every future interview?
The difficult thing with hiring bad candidates vs not hiring bad candidates is that you get to experience the bad hire every single day, so it is quite apparent that you made a mistake. With not hiring a good candidate, you don't really know what you were missing. You may just assume that the best applicant you reject is no better than your average employee. But there is a chance that that person would bring something to the table so unexpected that you can't even imagine the world that would have been had you hired them.
Apparently you hadn't been recruiting anyone in Europe. Once you've hired a bad candidate, it is very hard to get rid of it (borderline impossible in Scandinavia, for example).
I understand your point, and I'm not trying to be a troll.
Humans are more complex than that. I don't think you can assume that candidates will perform the same all the time. Sometimes an excellent candidate can perform badly for multiple reasons (e.g. nervousness, poor preparation, bad interviewer, personal problems, etc).
It seems to me, that rejecting a good candidate, and have him/her interview again after some time, if that candidate was a 'good-hire', then it would increase the chance of hiring him/her, since it is most likely they will prepare better, and know what to expect.
And why don't they count the cost of good candidates who simply drop out of their hiring pool entirely, because they can't be arsed to bone up on the easily gamed idiotic hiring process?
I'd never even apply to Google based on the stories I've heard, and I'm sure there are plenty of others in the same position who are even better at what they do than I am. There are just so many stories out there of how crappy the hiring process is that everyone who's any good must have heard about it. Some significant fraction will have said, "Yep, not for me."
Of course, people hired after their second interview are more likely to be mediocre, because people with more options are more likely to not re-interview (like mxcl https://twitter.com/mxcl/status/608687283869503488)
If only whiteboarding interview processes actually weeded out false positives.
In practice, they select for people with good memorization skills. If you can remember the details on a ten dozen different algorithms and data structures, you can pass one of these without having a single lick of creativity or skill.
I say this as an employee who has worked alongside many unskilled drones who made it past the algorithmic interview process.
I recently interviewed for a security engineering position at a startup, and while I got offered the job, the interview was quite silly.
The first thing I was asked to do was to write a lisp interpreter. It was a pretty trivial task, but it left me scratching my head since not only did it not tell them about whether I would be qualified at all, but I had explicitly avoided writing parsing code since I found ANTLR during undergrad.
In the end I wrote a basic stream abstraction and wrote my lisp interpreter on top of it with no real difficulty, but it was a completely stupid question. I told them on the third interview that they weren't asking me anything relevant and I still got another question later on about how to iterate over a tree in a specific order.
Part of the issue was that most of my interviewers were fresh out of college, it was literally their first job and they had no idea what to ask besides the kinds of questions they had been asked in their own interviews.
That last sentence is telling. Would you even consider working there if they had no idea what to do? I have worked at established companies where there is clearly nobody at the helm and it is frustrating.
I considered it for a bit, they were hiring their first security engineer, so it's understandable that they weren't very sure what they were hiring for.
In the end there were a lot of other issues that made me not take it, but not knowing what is necessary seems pretty typical wrt security. However, in those situations you want to have wide latitude to affect change because the thing they hire you to look at is probably not really where their problems are.
Infatuation with "coding competitions" and high flux intelligence tasks is often an anti signal.
Sure, what they demonstrate looks good at face value, but unless you attach a micromanager over their tasks, they'll constantly deviate to the new-shiny every week instead of staying on task for the long haul.
I know even YC has been bitten a few times by accepting people who are top-performers in very narrow contexts (e.g. people obsessed with winning coding competitions every single weekend) only to see they can't do anything coherent for more than a week at a time.
Just because someone is great in a narrow context doesn't mean they can carry that over the months and years needed to deeply iterate (and suffer through the low points) on successful projects.
I'd generalize your point a little more - for a given level of life success, intelligence will anti-correlate with other hire-ability skills like work ethic, diligence, attendance, emotional stability, professionalism, etc.
Basically, imagine that there's two ways to graduate college with a CS degree - you're either the kind of person who's smart enough to get things if you work your ass off for them, or you're a genius and you coast through. If you're a genius and you work your ass off, you get a doctorate or found a startup, and aren't in the candidate pool. So you wind up with a choice about what to compromise on: work ethic, or smarts.
Ideally, you compromise on whatever makes the least difference to your organization, compared to your competitors. Hiring smart/motivated people out of non-traditional backgrounds is a great option for this.
That's not what the interview process is like at all.
They're more interested in how you approach real world issues (the questions I got asked were conceivably real-life issues a company like Google would face with its products). If you can solve that issue by applying efficient and well understood data structures and algorithms, then that indicates you understand the problem space and solutions that may apply.
It's not like they get you in a room and ask you to draw a linked list or a binary tree.
I'll be starting at Google in a week and a half. I can't go into specifics, but I can say this:
I had two phone interviews - the first had quite a few questions about my prior experience (over a decade) and how to approach real-world problems from a high-level, including some that I had worked on at my current position. The second was primarily purely technical, but did have some more theoretical/project-management-y type of questions. In neither one was I required to write any code.
On-site: one interview was entirely whiteboard-coding (in a collaborative style), on a set of related generic problems. Next was one with a set of problems related to my experience. Third was almost entirely telling war stories. Last two were some generic graph-theory-type problems and some simple problems involving data structures relevant to the my area of experience.
As I hadn't interviewed for anything in 11 years, I did spend some time brushing up on my basic algorithmic theory, and essentially followed Steve Yegge's suggestions. I honestly think I probably could have done the vast majority of it without any studying, but spending time practicing solving problems on paper (with a countdown-timer as an artificial pressure-inducing device) certainly did help me, I think, in the on-site. No amount of memorization could have helped with most of it, aside from basic knowledge that anyone who's taken Data Structures should know.
I didn't find the process to be insulting at all; it was an enjoyable challenge. Everyone I talked with on the engineering side of things I can only say the best about, and so far as I can tell the team I'll be with has some really great people. I am a more senior developer, so perhaps I got a different experience than some.
When I interviewed with them, it seemed the interviewers had quite a large effect on how each session went. I had four great interviews, and two that were unpleasant. This had very little to do with how challenging the questions were and a lot more to do with the style of the interviewer.
My experience wasn't a real life type one more academic instead. My interviewer was really late and refused to let me finish my data structure implementation to optimize a single method instead. When we finished it was time for him to go. He wouldn't let me go back to finish my data structure as he didn't have the time.
I was later declined. Had a couple of friends at google ask the interviewer. He said I was declined for not finishing my data structure.
Their interviewing process was quite frustrating. I'm fine with getting declined but not under those circumstances.
I can assure you my story is accurate. The interviewers do not make decisions regarding the entire hiring event but they do give feedback regarding how you did in each interview. I was given poor feedback due to, according to him, a zero tolerance like policy where the data model for that particular question had to be complete and I was unable to complete it.
That was my first technical interview over the phone so it was pretty obvious it was related to that single interview. One of the Googlers I knew that checked the system said the same.
The important part of the "realistic" process is the part where you go from "the crawler's indexing is too slow" to "it might be because it allocates a lot of little bits of memory; it needs to first generate a BST of outbound link weights, and then functionally invert that into a second graph with a bunch of inbound link weights, and then inject all that into the PageRank matrix. You could probably save some time (though at the cost of memory-safety) by inverting the first BST in-place."
Once you get to that point, though, if they asked me "and how would you invert the tree in-place?" I'd answer "well, first I'd make sure we aren't just using a tree structure that already knows how to invert itself, or could easily be extended to know how. Then, assuming that didn't pan out, I'd google 'in-place invert binary tree' and read three or four solutions that already exist to burn the semantic structure of what I'm about to write into my head so that I don't subtly screw it up and have to fiddle with it for the next hour when I could be coding something else. Then I code it."
One of my interviewers at Google asked me to implement a balanced tree insert on the whiteboard in C. After doing so he immediately told me it wouldn't work. I searched futilely for a minute or two to find the problem before he condescendingly pointed out I had left out a semicolon on one of my statements...
Another in the same interview got in my face (this should be easy) when he thought it was taking me too long too multiply two numbers like 637 and 41 during an estimating problem. I understand the point of most of these questions is quick and dirty estimation (600 x 40), but seriously...
RE your second paragraph, it's easier than it sounds.
637 is an awkward number, we'll start by rounding up to 640:
640 * 2 = 1280, anyone in tech knows 2x that is 2560.
Make up the missing 3:
3 * 4 = 12, 2560 - 12 = 2548.
2548 * 10 = 25,480.
25,480 + 7 = 25,487
25,487 + 30 = 25,517
25,517 + 600 = 26,117.
That said, I did make a screw up when calculating it, and added the 637 before multiplying by 10 (which may have been the differentiating part of that question, combined with the intentional stress, to see how you cope with pressure? Although I didn't sleep last night, which might be the cause :D). I dunno if they wanted you to do it as quickly as 600/640 x 40 though, which obviously makes it much more difficult.
Afterthought, due to sleep deprived state, quicker way to solve:
Of course multiplying two numbers is easy: primary school children learn how to do it. Doing it in your head under pressure in an interview is entirely different.
> It's not like they get you in a room and ask you to draw a linked list or a binary tree.
And yet the OP was turned down for not being able to invert a binary tree (based on limited information, etc etc.)? What do you feel that is, if not a "whiteboard this algorithm" question?
As for many of the other companies who try and emulate Google - this is exactly what they do.
> There are tens of thousands of employees at Google. Most of them interview.
You say that as if it's a good thing. Do you expect that most people could be e.g. effective teachers, effective marketers, or effective taste-testers, with no previous experience?
That said, I recently heard about a Google interview where the interviewer asked the question and left the room for 30 minutes while the candidate white boarded.
Obviously he didn't give 1 shit about "how you approach the real world issues".
Ok, on the one hand, you should be able to answer the question with, "Get a standard-issue POSIX-conformant regex library. Next question!"
On the other hand, knowing that DFAs are equivalent to regexes are equivalent to NFAs is already a step of theoretical knowledge that any "Just let me write code" so-called "dev" is likely to complain about being asked for.
I was asked that question about 7 years after reading Mastering Regular Expressions and about 10 years after taking Automata in college. It so happens that I am an expert in regex, have a patent in the field, regularly answer stackoverflow regex questions for fun, etc. Unfortunately, the question involved me actually drawing a DFA, trying to remember the notation, etc. At this point I don't even remember the particular question, only that I couldn't remember how to draw or read the diagrams.
The interviewer told me this was a favorite question of his. Also I don't think I presented myself as a regex expert to them. Anyhow, I'm an expert in regex like a DBA would be an expert in SQL or particular databases. The interview was for an ordinary software engineer role.
Regardless, I only meant to make the point that I disagree with the comment "They're more interested in how you approach real world issues" and that this statement is false: "It's not like they get you in a room and ask you to draw a linked list or a binary tree"
edit: by "expert" I just mean something I'm particularly skilled in, not something I'm an authority in from theory to practice.
I think what Google tries to find are the engineers with creativity, skill and ability to work intelligently at a problem which they haven't seen before.
It's important to know how a person approaches a hard new problem and how easily or not they give up. I'd guess that a person who didn't solve a hard problem, but got close enough by displaying creativity, curiousness and a solid thought process might have an edge over a person that implemented a familiar algorithm with ease but got a little flustered, and didn't show much creativity or inquisitiveness, when challenged with something new.
I think you want to think that way, that whiteboard interviews have a shred of competency about them. I bet you dollars to doughnuts though 9/10 times a company will hire someone who can regurgitate algorithms rather than an eccentric who can't get them right but thinks really cleverly.
It seems this criticism is of the interviewer making a poor decision based on missed/incorrect observations, not with the whiteboard interview itself.
Whiteboard interviews might miss eccentrics who can't express their clever thoughts verbally or visually, but it can depend on how tuned in the interviewer is to these eccentricities or how much value their team puts on expressive skills.
Algorithm/data structure interviews, whiteboard or not, also have the added benefit of being exceptionally good at eliminating candidates that:
I've never answered it wrong and gotten hired for my "thought process" even if I was close. I had to fart around on my own for another ~15 minutes to fix my bugs to make it work. But it seems they just want a guy who memorized it and pretends to discover the perfect solution on the first try.
On two occasions I have "re-interviewed" current or past developers who wanted to move up from working on extremely simple low-complexity code / sysadmin work, and start working on our flagship project. In both cases the team leads and I didn't think these people were qualified to do so, but the candidates were really convinced they were. These were employees who had come into the company through non-standard channels, so hadn't experienced our normal interviews.
I had them attempt to solve the problems on a whiteboard that I would give to our intern candidates. Both totally bombed hard, and our other "re-interviewers" had the same result when they asked some design-related and algorithmic related questions.
This wasn't "regurgitating algorithms". It was basic problem-solving and software design. From their track record of development we didn't believe them to be qualified (despite their own opinion of themselves), and they only confirmed it by totally bombing the "re-interview" process (which was identical to our normal interview process).
Nothing you said actually means they could not have done the job. You basically said, we gave them a test and they failed so clearly the test was a good idea.
The only way to validate your highering process is to randomly accept people you would otherwise fail. Anything else is meaningless.
Say I pass the invert a tree question. Does that prove I am good? That I can design a product, listen to customer requirements, come up with ideas that elude others, read a research paper and turn it into something commercial, solder a circuit board, make a schedule, mentor junior engineers, write documentation? No. All inverting a tree tells them is I studied trees recently, and/or I'm at least superficially clever.
Google lets go of plenty of people. They aren't making perfect hires.
They have publicly admitted there's no correlation between their hiring process and outcomes. I'm not even clear [why] we're still debating this.
Other than to wonder why they still haven't done anything about it. (I interviewed there ~8 years ago, pre Android when they were looking for deep mobile expertise and got flunked out on a similar problem. I am not at all interested in working for them now)
To show that there is a correlation you have to quantify both hiring process evaluation and job performance. Since both are elusive quantities that are hard to quantify, showing there is no correlation may just mean that measurements are done poorly. E.g. when job success is measured by boss satisfaction.
But do you? How catastrophical would it be if Google accidentally hired not the top 0.001% but instead somebody from top 0.01%? I mean, same people that build everything else around us and the world still didn't collapse yet? I get it, you have to have standards, and you don't want to waste time and money on somebody that can't pass fizzbuzz test. But once you've moved past that - do you need to obsess so hard on ranking and quantifying it? There are probably some positions in Google that require people to invert binary trees (speaking figuratively) day in and day out, but my decades of experience in the industry shows that most positions aren't like that. I've not worked for google but I'm pretty sure it's also this way there. If you hire smart and competent people, it's OK even if they don't reverse a binary tree here or there.
Yes, exactly, since it's Google, they can afford to have strong filters which will definitely filter out a bad hire, may be at the expense of rejecting a good candidate. Their policy might expect good candidate to get absorbed in Google one way or other, but they definitely want to filter a bad hire. However, in the recent years, I have seen this to be not working as expected. Still, I don't think it matters to Google as an big organization.
Good teams have a balanced skill sets I've found. You need some uber-algorithm guys who can pass questions like this, but you also need engineers who can fill in all the rest of the tasks that make up a production worthy system. I have not found that the uber-algorithm guys always pay attention to detail, put in good comments, think of all the edge cases, communicate well with outside teams, etc.
Those types of engineers are what you consider a "bad hire", and won't be working at Google.
> you also need engineers who can fill in all the rest of the tasks that make up a production worthy system. I have not found that the uber-algorithm guys always pay attention to detail, put in good comments, think of all the edge cases, communicate well with outside teams, etc.
What's to keep a good engineer like you describe from buying a couple of books and spending a couple of months working on becoming the type of person that can do well on these questions?
> What's to keep a good engineer like you describe from buying a couple of books and spending a couple of months working on becoming the type of person that can do well on these questions?
Google doesn't make everyone multi-millionaires anymore. If you want me to jump through a bunch of hoops for you, you'd better have a pretty solid case for what the outcome will be.
By actively selecting against people who don't want to answer bullshit questions, they're definitely turning away creative, talented people. They're selecting for different people instead. Perhaps that's what Google needs to feed the machine now that the company is mature, so I won't necessarily fault them for it. But that filtering or selection IS happening, even if they don't think it is.
I'd rather flip it around - what's the reason for me to do that? I don't see them offering obscene piles of money anymore, so they don't get to make up ridiculous hoops and demand everybody jump through them. This stuff is rarely useful in the real world, and when it is, you can grab something off the shelf, or look it up then. I'd rather spend time learning the other stuff on that list, that's useful to companies that have their priorities straight.
There is a glut of interesting opportunities out there. For me personally, there are just many more interesting ways to spend my time than revising the CS classes I took almost 15 years ago.
In the long run, this could affect Google, but, then again, if they change their hiring policies for better even after long 5 years, the best talent that might had been rejected before would still go to work at Google. So, still I don't think Google really consider it as an issue for them. Though, I have been burned once myself by them.
I don't consider myself "top talent", so I won't comment on my own experience through the interview process. But I do know that I don't want to work for google. Hell, I'm happier doing my own thing.
I also know a few "top talent" engineers that rejected the google offer because they didn't like the process and the people they met. I know even more ex-Googlers that quit within the year of starting there.
It's not an issue for them? From where I'm looking, most of the experienced top talent (that it's not working at google already) is keeping away from it, in greener pastures and most won't look back.
Yep, google is not going anywhere anytime soon, but I'd be just a bit concerned if I were them.
Google is not the Google it was. I mean I still consider that working for Google is attractive, but things I get from tens of my friends who work there are to say the best....mixed. Google might still have the vigor and excitement at its core, but unless you are an already exceptional engineer, the work you get assigned there, often times, ranges from bland to boring.
Like I said, Google is still an amazing company, but it is by no means the only one out there. Talents won't come to work for it just because it is Google, without great compensation and interesting projects.
>>However, in the recent years, I have seen this to be not working as expected.
It never worked, what do you think happens when you start asking these kind of questions in interviews? People just go out and spend insane amount of time memorizing solutions to interview questions on the internet.
They continue the same after you have hired them. So their on the job productivity remains low. Please note, they need to prepare for their next job in the current one.
So the only thing this results is in people spending insane amount of time preparing interviews doing very little work in their day jobs.
I'd rather spend a year learning Python or Go or whatever, instead of spending a year practicing interview questions.
I never realized that some people put as much effort into practicing interview questions as I put into learning how to actually write software.
Does this mean technical interviews, even a structured interview, are completely useless? I.e., if you follow an interview script, eventually candidates find out about your script, and then it's useless.
This cracks me up a bit. I don't have anything against whiteboarding, even though I find it meaningless. With this you more often than not have false positives.
And if Google had so many resumes they have to filter through, then why do they have all these recruiters literally spamming you to join? they have recruiting issues like many others tech companies, but the whiteboard is really a bad filter. I've seen very much capable,competent and experienced engineers being denied just because they didn't spend too much time learning cute algorithms or coding in the whiteboard, in front of interviewers sometimes more focused in showing off how smart they are.
"Google's perspective" shouldn't be a monolithic thing. Sure if you're sorting through candidates that were dropped on you then this process makes sense. But if you've discovered someone that by reputation and demonstrated code base has already excelled, and you're actively chasing them, then the interview process for that individual has to change. It's more about seeing if they'll thrive in your environment, and convincing them that you're an inevitable part of their career.
It's true that evaluating candidates is a hard problem, but you'd think that at the very least they would try to evaluate them based on skills that will actually be used in the job. Judging programmers based on their whiteboard skills is like judging a ballet dancer based on how eloquently he can describe his training routine.
Well, my suggestion would be to use Page Rank to sift through the resumes. If you are respected by respected computer programmers then your "score" goes up. They could measure that by mentions in a blog, twitter "conversations", Github commits on the same project, etc.
"Hmmm... it looks like this guy really gets around. He's on the same project at GoodProgrammer A and GoodProgrammer B. He also tweets about functional programming and was mentioned in GoodProgrammer C's latest blog."
People have previously mentioned hallway discussions at Google routinely devolve into bragging about SAT scores and GPAs.
Nothing says Google must accept sub-par technical people, but Google fails to realize you don't have to be the top 0.0001% in absolute algorithms intelligence to do great work.
Google's basic theme seems to be "you must be low latency ultra smart in our narrow interest areas" to pass an interview (or well-connected in other ways to lower scrutiny), but Google's main problem these days isn't lack of smart people — it's lack of projects with meaning and lack of how to organize smart people into anything resembling a coherent world-moving force.
In The Plex is great read for insight into their entire process. Basically, Google is run by people who have never been told "no" to anything in their lives, so they continue to think they are the best at everything until reality forces them to reconsider their delusional assumptions ("montessori naivety").
In The Plex is great read for insight into their entire process. Basically, Google is run by people who have never been told "no" to anything in their lives, so they continue to think they are the best at everything until reality forces them to reconsider their delusional assumptions ("montessori naivety").
Which is unfortunate, because objectively Google doesn't actually have a very impressive track record of creating successful products and services given its size and resources. It has a goose that lays golden eggs (the on-line advertising business), two strong on-line services (search and mail, both over a decade old) and two well-established platforms (Android and Chrome, both nearly seven years old). Other than those, most things Google tries seem to land somewhere between unremarkable and complete failure, with Google+ surely the most obvious example of the latter. They also seem to be developing an unenviable reputation for killing things off as a result, which is going to make it more difficult for them to succeed with future endeavours. Whatever workforce their hiring process is producing for them, it doesn't seem to be one that is very good at creating successful new projects.
Triple up-vote for this. That said, As said before, I would say if they get more applications than anyone else, their pool of possibilities is larger, so, even if their hiring process is sub par, it's still possible to get more better people. They have a lot of good people and quite a few really good people. And, likely more of those really good people. Someone has done the numbers.
It's a question of efficiency and maybe one of those smart ones is optimising it. It's quite possible that the process that looks shit on the outside is producing the right numbers.
It seems we have different definitions of "excellent".
Google Maps is certainly widespread, but it's hardly the only on-line mapping service available. They keep messing around with the UI, often not for the better. I often find the data itself and the live route planning/traffic news to be inaccurate.
Maybe it's better for those in the US, but here in the UK it's exactly what I meant by unremarkable. The local data from OpenStreetMap seems to be just as good for general mapping purposes and sometimes more accurate and/or current, and for car journey planning and real-time traffic news the dedicated satnav devices still seem to do a far better job.
And Google Maps was built from acquisitions like Where 2 and Keyhole. Their data processing has improved vastly, but the product seems mostly unchanged.
> People have previously mentioned hallway discussions at Google routinely devolve into bragging about SAT scores and GPAs.
I've been at Google for 3.5 years (in the NY office) and have never heard this. I don't think I've ever even seen anyone wear a college t-shirt/sweater.
> These aren't Googlers I'm talking about, except the boyfriend.... remember that most of these conversations happened when I was actually in college (a few even before then, while applying to schools), and I just have a really long memory
Sounds like someone who's just into remembering people's SAT scores?
I worked in one of the Mountain View offices for a year and a half with people who had never even been to college, and never heard that. Doesn't mean it never happened but the idea that it's even close to common is just laughable.
I do wear my college t-shirt though. My graduating CS class was three people, so have to have a little pride there!
My friend worked at Google a few years ago (in a non-engineering role). After his in-person interviews, the recruiter called him back to ask him how many hours he worked during college (to pay his way), so as to excuse is less than 4.0 GPA to the hiring review board. He had completed his undergrad degree was 20 years in a subject not directly related to the Google position. He since had an relevant MS degree and real-world work experience, but that undergrad GPA was apparently still really important.
> Nothing says Google must accept sub-par technical people, but Google fails to realize you don't have to be the top 0.0001% in absolute algorithms intelligence to do great work.
But if you're Google, why wouldn't you aim for the top 0.0001%? It's not as though you have a shortage of candidates.
Because there's no such thing as "the top 0.0001%". Software development is a collection of talents with multiple basis vectors, not a single linear talent you can stack rank.
And even if there were such a thing, you're not going to find the top n% by asking questions from an undergraduate textbook.
And you're certainly not going to find them by pissing off people with proven practical achievements. Acting like that is evidence of anti-intelligence, not a sign of being smarter than everyone else.
Do the math: 0.0001% of 7 billion is: 7,000. How many employees does Google have? Not to mention the fact that of those 7 billion, probably at least 6.9 billion have no appreciable training in software development, which would leave you with 100 candidates.
> Google is really interested in hiring the best engineers but rather a specific type of engineer.
I think you made an excellent point. It seems that outside a tiny minority of world-class experts, Google is more interested in hiring interchangeable workers.
Exactly. I see it as a tacit admission that you are effectively a cog in a wheel if those are the kinds of questions they are asking.
I know some people get frustrated by those kinds of questions, but I'd just see it as a reverse weeding out process. If you still want to work at (insert company here), just build something valuable and have them acquire you. At that point they'll acqui-hire you and you'll leap frog all the people who got in because they had Knuth books memorized.
Really? Being able to solve basic CS101 problems requires some sort of privileged upbringing? I grew up poor - my family lived off food stamps for a while. This stuff is easy. I find it disgusting the level of entitlement swept throughout this thread. Oh Google HAD to give you that job, it was yours and they stole it by testing basic skills.
You know, there is something to be said for someone who has a good product sense, but not good developer skills. They sound very valuable, so why are they applying to a developer position? Just because you are unqualified for one job, doesn't mean you deserved that job, or that you're worthless.
I suspect the issue is that programmers right out of college are strong on testable skills (like how to do X to a linked list) and weak on experience-based skills that are harder to test (like how to keep projects under budget or how to manage technical debt). Whereas veterans naturally tend to be the opposite. And hiring processes naturally tend towards what's easier to test.
As such, the best approach to these interviews might be to say, "Look, I could take a stab at that algo question, but I haven't touched that stuff since school, and if I needed it I'd google it. It might be more productive if we talked about what I've been doing day-to-day for the last ten years to ship successful products, because that's the main value I can bring to your company."
I aced some data analyst interviews fresh out of grad school. A lot of the questions were excel and probability questions, which I used to teach. Now, I have more experience and knowledge, but I would be much slower in those interviews.
>not so sure Google is really interested in hiring the best engineers but rather a specific type of engineer.
Any mature company should focus on finding those specific candidates - the ones who fit the company and specific team, on several dimensions beyond standard skillsrequired for the job. That's someone who will be called the best candidate in that specific company. And it's absolutely natural that s/he won't be considered "best" insome other company.
I interviewed @ google several years ago as part of some due-diligence they were doing to decide if they wanted to acquire the startup I was working at, [redacted].
It was several hours, with many interviewers one at a time, of academic exercises. Some I could see were relevant to some of Google's projects, but that wasn't the emphasis- it was very much a "how much do you remember from CS courses?" (with a few exceptions- one had me explain why a snippet of C would crash- something more what I would expect a senior engineer to know that was more practical-oriented).
I _think_ I did pretty well (again, it was part of a bigger due-diligence-- another story altogether, so I never heard), but in the end I was struck by the fact that (a) not a single interviewer had looked at or asked about any of my publicly available work, and (b) not a single one even asked something along the lines of "what are your strengths / what do you think your contribution would be" (or weaknesses, for that matter, but that seems less important).
Of course, one of the other guys I worked with lucked out and got interviewed by one of the paypal-mafia, and ended up with a reportedly very stimulating hour long discussion about actual work, strengths, and actual situations that one would hypothetically need to handle at google.
Your previous work is what gets you in the door, but interviews at Google are explicitly structured so that you need to demonstrate competence right then and there.
And ye olde "what's your greatest strength/weakness" question sounds like a terrible way to judge an engineer.
Makes sense for google. And it's not that I didn't appreciate doing all that (and I did do well), it just struck me how they knew nothing else about me afterward except that I could answer academic questions, and how different that was from the startup world- where, at the very least, you want to know what they're passionate about and that they are productive.
Asking an engineer what their strengths are and then to demonstrate or explain in concrete terms is how you find out their unique value proposition. I'm not talking about "I'm a hard worker" type garbage, I mean "my strongest experience is in systems programming on Linux." "How so?" "For example, I wrote a specialized TCP kernel module that works like this..."
Again, not saying that should be the whole interview- but there are some non-academic skills and experience (like managing and leading a critical system used by thousands of people) that are worth more than a few weaknesses in problems that were already solved years ago and that are easily looked up. It's how you should judge people looking to get into a graduate program, but falls short if you're trying to build teams of awesome fury that need more broad experience and even (gasp) unique experience that you don't have a ready-made question for ;-)
For broad-spectrum roles like software engineers, Google hires people first and then figures out what team they should join later. If you want a specialist role, then you'll need to find one on their jobs page.
> not a single interviewer had looked at or asked about any of my publicly available work
It could be dangerous. Let's say you explain some interesting techniques you used on your GPL'd software and a Google product starts doing something suspiciously like what you showed them. You may even sue and end up making their source public after some litigation.
You can't make anyone making source code public, only stopping them from distributing copies of code which you have authored. Copyright and patents works very differently, and its quite hard (impossible) to copy source code wholesale from a interview by accident.
If they did makes copies of an applicants copyrighted work and distribute it, then they could get sued. Some companies has been know to explicitly ask applicants to bring source code from previous work places (also called industry espionage), and those been clearly intended crimes and not something that just happened to company during the process of an interview.
Replying to myself- too late to edit. Some other comments in here lead me to believe there would be experience-related questions ideally these days. Again, my experience was a few years ago :-)
"How much do you want to work here?" is probably a better predictor of your performance on the job than your experience is. I don't see the problem with that: IMHO, the fresh grad who really really wants to be there should be hired over the experienced veteran who's just putting in time. I'd like to think that an experienced veteran who really really wants to be there would also get the job, though, over both of the other candidates.
> At a certain point, your resume should speak for itself. The fact that experienced engineers with impressive resumes are put through these types of interviews is insulting and frustrating to the interviewees.
Only to bad interviewees.
You'd be shocked to hear the number of impressive résumés I have seen who turned out to be absolute no-hires after 20mn on a white board. By "absolute no hire", I mean the kind where some of the interviewers tell me in the debriefing "I will quit if we hire this guy".
Not all excellent engineers have impressive résumés but they all share a common characteristic in my experience: they all really enjoyed working on the problems I asked them to solve on the white board. I even had some of them be disappointed when we ran out of time because we were having interesting conversations about the problem.
> You'd be shocked to hear the number of impressive résumés I have seen who turned out to be absolute no-hires after 20mn on a white board.
Hmm, at least in my experience it's really quick and easy to tell if someone is bullshitting about their resume or not by simply talking about the different projects they worked on. I know a few developers who didn't go to a 4 year school so they don't know all of the correct terms or algorithms (one even interviewed for Google and failed) but I would hire them in a heart beat because they are really awesome.
I still feel some sort of pair-programming exercise or something similar is better than simply trying to solve problems on a white board. Unless you're in a SCIF you're going to have access to the internet to refresh your memory, etc so I've never found utility in asking academic questions. In fact I've had a few people who could pass the academic questions but couldn't build an application to save their life so I don't even bother anymore.
>By "absolute no hire", I mean the kind where some of the interviewers tell me in the debriefing "I will quit if we hire this guy".
That's just a self-reinforcing selection bias - you hire people that are good at these kinds of problems/enjoy them - you filter out the rest by default and it becomes ingrained in the culture - it doesn't say anything about the suitability of the candidates you filtered out. If you said "we hired a guy who had a great resume but he couldn't do algorithmic problems, turned out to be a bad fit for the company" - that would still be a single data point but at least it would be relevant.
I would agree with what you said if you replaced resume with reputation. Anyone can fill out a nice resume.
One of the challenges with Google's interview process [0] is they are very worried about hiring mistakes, and also know that the process is very noisy. As such, they interview a tremendous amount of people for hiring. In addition, they've found it more efficient to just keep a high bar (missing many good hires to keep out 1 bad hire) than to conduct more than 5 interviews.
The one thing I've wonder is - if they want the top 100 computer scientists in the world to work for them, will this process produce that? Or is there another process for them?
I think the most effective way to get a job there is to build a small company and get acquihired.
[0] I don't work at Google, but I've researched the process in anticipation of an interview.
Well I know one mistake they may have made in hiring. Guy I used to work with had a totally unclear design process and couldn't complete his work. He would design flawed algorithms that made no sense to anyway and then he asked for lead a project which he dug himself into a hole and had me pulled in to help him and then abandoned me and left the company. Spent 3 months cleaning up his mess. I wonder how he is working out for them...
Ah, they are very unlikely to get the top 100, but very likely to get e.g. 100 of the top 200? Now I think I understand the trade off they are going for (e.g. Getting 100 of the top 200 is more efficient and they are "good enough")
Eh I don't know that they're getting 100 out of the top 200 either. At some point Google's reputation for an insane hiring process causes the best computer scientists to not even bother applying because it's a waste of their time. Those people have many, many options that would pay them handsomely for their talents.
Look at it this way. If you're a cream of the crop programmer and you're looking for a job, Google has the most drawn out and arduous hiring process. Even if you apply there and start the interview process it often takes a lot longer than the industry average to get to an offer. In the meantime a multitude of other companies have probably already determined that you are an excellent programmer and tried their best to snap you up.
Google seems to be happy with the tradeoff they're making but I'm pretty sure their process is not optimized for getting the best people in the field.
It's also stressful. To be fair the tree problem he got was easy, but brain freeze can kill you. I flubbed a trivial problem at a FB interview that still haunts me today.
Often it's completely free-form too. So you have to pick the language, the data model, the data structures, the traversal mechanism (recursive vs. iterative), state to track (allow duplicates? prevent infinite loops?) and how to test the example all at once before even getting _started_ with their problem description.
Many of the interviews could actually work if they broke the process down into 8 questions instead of one (how would you represent a tree? a binary tree? how would you find node X? how would you move node X to the other side of the tree?). But, the ability to accommodate understanding takes empathy and human connection, not being a jerk programmer on interview duty who is browsing HN while listing to the interviewee mumble on your speakerphone.
Never underestimate the lack of ability in the interviewer and not the interviewee. Except, the company always believes itself to be more competent than outsiders, so you're ultimately left with being condescended to unnecessarily.
Ha, I totally messed up a question for a front end interview today that could have easily been solved with timeouts. I didn't realize timeouts had resets as I haven't used them extensively and bam I'm trying to essentially design a reset for a generic counter. Boy I felt ridiculous. As soon as he said use timeout reset it's immediately obvious what I had to do. I'm not a bad programmer but between brain freeze and whiteboarding, and interviewing day after day for a few weeks, it just didn't occur. Killed me and I hope I don't miss out on an offer from a company I was actually really interested in because of some bullshit brainfart on something I'm not particularly well versed in.
Whatever! They asked me back but it sure is a frustrating feeling even if I tend not to appear flustered until I walk out, introspect, and realize what a doofus I was.
They don't just stop at binary trees. Think of it more like memorizing hundreds of magic tricks that could be done using a deck of cards. The tree is more analogous to those deck of cards. The data structure remains constant but the tricks you can do keep changing.
I know of situations where candidates have been grilled whole full hours on tree traversal methods, and then asked to code up perfectly working solutions for them.
Do these people seriously think people who can invent 50 years of CS research by sheer analysis over an hour on the whiteboard, will be ready to work for them to write shell scripts? If not you are only testing memory skills.
In my academic days I used to refer to say "physics-standing-up" is a different subject from "physics-sitting-down" and needs to be studied separately. Some years after I left academia I had the good fortune to get in contact with the person who introduced me to acting at a very young age and thank her for her contribution to my academic career. About 30% of my success was due to my comfort working in front of an audience--even a panel of examiners is "an audience"--and I had a huge advantage over my peers because of it.
One thing that I'd recommend to anyone seeking any kind of job is taking some improv courses--they are pretty common in major centers and even some smaller towns, and can really improve your ability to deal with this kind of nonsensical interview situation, which tests a bunch of skills that have nothing to do with your ability to actually do the job.
I'll disagree that these questions require weeks of preparation. I think it's more just unfair because they favor a certain type of person: those whose interests align more strongly with the theory and algorithms and less with building things.
Of course, the irony is that for vast majority of the people they hire, most of the work they do will be building and not much algorithms work. So then what will happen is either you have people who are great at building things and force them to suffer through the interviews or you have people who are great at algorithms but not that interested in programming.
Of course, you might also have people who are great at both, but I wonder how many of them there are..
I don't necessarily agree though, I think if you can reason about a problem, you are infinitely more valuable than someone who memorizes problems and just applies them blindly.
Correct me if I'm wrong, but a binary search tree is a very simple data structure. It has rules, and just by it's definition, and the definition of "inverted", you should be able to come up with an iterative solution, even if it's just walking the tree and re-creating it.
It may not be the best, but it's a solution, it's a starting point.
But if you're interviewing for a job solving problems, shouldn't you be able to solve problems? Even if they aren't the best solution? Genuinely asking here, as this approach seems to have a 85-90% success rate for me.
It depends how your interviewer is ranking/scoring you. One interviewer might be impressed with your ability to talk through and solve the problem from first principles, where another might ding you for not instantly knowing the answer. The former type of interviewer would be happy to give you hints and prompts, the second is a stone wall. I try to be the first -- it helps assess how a candidate responds to coaching too! -- but there's a spectrum between "nudging" and "telling them the answer straight out".
I've definitely seen interviewers (both by reading other people's feedback, and in one rare case, being an interviewee) who expected a "correct" answer and wouldn't work with the candidate to get there if they didn't already know it.
See also: overused brainteasers. I remember being asked the 5 litre / 3 litre / get 4 litres problem for my Microsoft intern interview, and honestly told them that I already knew the answer, so I wouldn't be "solving" it, but here it is anyway...
I've seen the same thing, but I tried to account for it in the 10-15% of cases where I haven't had success.
Once I notice that it's not really going well, I try to politely as possible suggest that maybe we adjourn, so as to not waste each others' time. That hasn't really ever backfired, and actually turned one rejection into an offer, but I find it's polite to respect everyones' time.
Is a binary search tree a problem that the Dev would have needed to solve? In over 5 years as a dev, I have never once needed to solve a binary search tree. Why not ask the candidate how to perform a kidney transplant.. That's solving a problem too. White boarding algorithms is not relevant to the majority of Dev jobs. It's related, but not particularly valuable as a screener.
I'm not sure how asking someone who's going to be working in computer programming something about computer science is equal to asking someone who's going to be working in computer programming something about medicine. Those two things aren't equivalent. Could you perhaps clarify?
Engineering is the use of knowledge/science to solve problems, right? So why is asking someone to use knowledge to solve a problem not applicable to the hiring process?
To be clear, this isn't the same question as "why are manhole covers round", or "how would you move a mountain".
The act of creating a, say a self balancing binary search tree, is something you should know about. If you ever have to implement it, you go look it up. Engineers are exactly the same - you would rather have one that know where to look up all the stuff, or search for information, than one that can regurgitate the entirety of his education theory.
Typing "how to reverse a binary search tree" into google is not the same thing as solving a problem.
If these problems were so reliant on individual problem solving ability, there wouldn't be dozens of books with reams of answers to these questions on the market. These questions test memorization and nothing else.
I mean, that's a little strong. My undergrad degree is in math, so I would be thrilled if I was sitting in an interview and was being asked questions like that instead of stuff about how computers actually work.
Google said early on they want to hire the best and brightest and that was the most important. We got the mantra that passing over a good person is better than hiring a bad one. And they were right. But they're a huge company now. They, and other gigantic tech companies are always complaining they can't find 'good people'.
But how do you make this philosophy scale? Do you say, "This person is smart but lacking some skills so we'll invest in them"?
Smart people will learn what they need to to do a job if they're interested in it.
Google has decided that the only way to tell if someone is smart is by asking academic questions.
Even though anyone with half a brain knows that there are other ways to gauge how smart a person is... such as experience.
Does anyone really think Max can't look up how to invert a binary tree if he needs to do that? He obviously knows how to read and learn. So why is that a reason to reject someone?
Google is filled with smart people.. but their interview process is pretty stupid.
> Google is filled with smart people.. but their interview process is pretty stupid.
Google is filled with people who can solve Data Structure / Algorithm questions in 50 minutes on a whiteboard. Their interview process selects for people whose past experiences led them to being able to solve Data Structure / Algorithm questions in 50 minutes on a whiteboard.
The extent to which solving Data Structure / Algorithm questions in 50 minutes has anything to do with building maintainable production software, developing talent, a good working environment, or even this mysterious concept known as "being smart" is up for debate.
> Google is filled with people who can solve Data Structure / Algorithm questions in 50 minutes on a whiteboard.
I wonder about that, actually. I mean sure, right after they're hired, I bet they're GREAT at doing whiteboard problems. But then after a few years of real work, I bet they've coded about as many algorithms from scratch as the guy from Unknown, Inc. (which is approximately two) and they probably wouldn't even pass another Google interview without a boatload of practice.
> The gigantic tech companies are always complaining they can't find 'good people'.
The "can't find good people" argument can usually be solved either 1) raising the pay, or 2) increasing the labor pool (i.e. H1B, etc). Raising the pay is what is done in finance because it's harder to make the argument that the required "skills" are lacking in the labor pool. To enter investment banking, and then private equity, it often only requires a B.A. in French literature and willingness to work 90 hours.
The "supply problem of engineers" is the great lie of our industry of our time.
I guarantee if dev positions paid as much as medicine or ibanking you'd have more devs. Not necessarily really good devs, but certainly people who pass the Ivy + 1600 SAT sniff test and can memorize the SAT book of Google interview questions.
I wonder about that. What percentage of their employees are recent college graduates? From all of these anecdotes, it seems like their process is optimized to select cheap recent CS grads.
I agree. I feel the overpowering urge, when asked this sort of basic interview question, to shout back "WHY?". What does inverting a binary tree have to do with software engineering, at all?
Google (and other interviewers) might as well be asking me to name the last 1000 digits of PI. The Kafkaesquen ability to recite how to do the task correctly reflects nothing about the candidate, except the strength of their memory.
It boggles the mind that the field of computer science does not have any standard way to detect the competency of a person. It is extremely discongruous for a company with as many smart people as Google to have such an absurd hiring policy.
At my job, our interview guidelines explicitly say that we don't expect non-graduates to be able to whip out fancy data structures. We want to see that the person can program, how he thinks about problems and system design, and how they deal with being stuck. I've seen people with impressive CVs and PhDs that were unable to solve very simple programming tasks.
Google interviewers are strictly told to ask these data structures and algorithms questions, it doesn't matter to them if they are relevant for future work. I have even read reviews by PhD candidates where even they were asked these questions in the interviews. When I interviewed with Google last year, I remember that the interviewers had a blueprint of the type of questions they were allowed to ask. They hardly concentrate on your resume during the interviews.
At what point during one's career should one simply be hired at face value and be exempt from interviewing? After 10 years? 20? 30?
"I see your resume says you've been writing software for 30 years, and wrote this really well known tool. Wellllll....That's good enough for me. You're hired!"
Background: I worked at Google for 5 years and did a bunch of interviewing for them.
I don't support skipping interviews even for good candidates, but if you were going to it would work the other way around -- you'd value recent accomplishments over more distant ones.
People don't understand that cutting out folks who had their name attached to some big successes but can't solve a toy problem in an interview is usually a really good thing. It's a strength of the system more often than a weakness. You wouldn't believe the number of senior candidates I interviewed who had really impressive resumes but were clearly checked out from doing real technical work -- folks who obviously no longer lived close to the code and either were very rusty or just no longer had the temperament to spend time thinking through a slightly tricky piece of code.
One of the reasons not to skip an interview for a candidate you're confident is good is that a good interview can be a great sales pitch to come work for your company.
I've interviewed and worked with 30 year programmers and for some of them it was obvious could not program and their resumes were concocted. I would never rely on a resume.
"Self-assessed" as a resume speaking for the candidate.
Without running a `git blame`, the interviewer cannot say how much of the code is from the author and how much is contributed by someone else and even then asking questions is an important part of any interview.
Aren't we trying to hire clever programmers? Isn't that the point?
If a guy has some kind of a git-mutating program that makes it seem like he writes good stuff, he's probably a good programmer to be able to write such a thing.
I guess everything I do is 'solving problems'. That's pretty open-ended. Maybe 'describe a situation where things didn't work the way they were supposed to/documented to work. How did you resolve it?' or something like that.
Another way of looking at it is that they do these kinds of interviews particularly because they do want to hire fresh grads (and ideally, the best 5% of those) over experienced programmers. And if that's the case, then it's not really flawed, it's doing what it's supposed to do.
Actually, Max himself said that it "Felt like they would have preferred a fresh comp sci grad."[1]
Investment banks recruit bright young graduates, putting them through a tough selection process (c.f. all those hard-working interns we keep reading about), and select the ones who perform/conform the best. The rationale for bringing in young (as opposed to experienced) people is that they'll work their asses off, can be moulded to fit the culture (i.e. independent thinkers aren't necessarily sought), and (perhaps most importantly) are too junior to present a threat to the incumbents.
If Google's going down the same path, it may be a sign that their corporate culture is solidifying, which would be a bad sign, IMHO.
I think this is probably overall true. I'm a bit cynical, maybe, but the set of tech used by developers here at Google does not necessarily intersect with what is used in the rest of the industry. So much is proprietary, and best practices are heavily codified and well understood. So in reality I suspect that new grads are as valuable or more valuable to Google as experienced engineers since it's unlikely the specific technologies an experienced engineer has under their belt will be specifically applicable here.
For the top elite of experienced engineers that may not be the case, as they will come with some overall good analytical skills that Google wants. But for most in our industry, what we have learned over the years is how to be good in various sets of tech around our platforms of choice... and that is mostly irrelevant to Google, as they have their own platforms.
It always surprises me when people like Pacino or De Niro talk about going for auditions - I'd imagine they had a large enough body of work that directors would know what they can do.
But then it occurs to me that they won't have seen them in -this- role acting -this- way for -this- film/play/whatnot and that's what the director needs. And given the amount of money and risk involved, it's probably wise to check first...
That's true. However In Google defense, they tell you upfront what kind of interviews you're going to get and their process it's pretty well known anyway (with all its flaws).
I went through the same pain myself a few years ago and I'm not disagreeing that the chap should have just got the job there, but their process is well known and perhaps he just had the wrong expectations.
(I'm at Microsoft, FWIW, every engineering team here does hiring their own way)
My team does whiteboard questions, but we try to keep them practical.
Typically they are the types of problems that we'd expect new engineers to have to look into on their first day. Oftentimes the questions are less "come up and an answer" and more "let's explore this problem domain and see what we can uncover."
As an example, the interviewer will write up on the board a data structure that is too big and has redundant fields in it, and start a discussion about how to optimize that structure. The optimizations go more and more in-depth as the candidate gains experience thinking within the problem space.
Another problem, the interviewer will give an example of a synchronous operation and work with the candidate to break it up into a set of asynchronous calls.
My team tries to avoid "puzzle questions with 1 correct answer".
I do ask simple data structure questions just to test basic programming skills. On occasion someone does make it through who cannot even write a FOR loop.
My standard go to question is "Please do an in order print of this already sorted binary tree."
It is 5 lines of code, and ~80% of candidates who haven't seen a tree in a decade can still figure it out.
IMH the goal isn't to try and trick people, it is to try and see if people can hold a conversation about a technical topic and make valuable contributions to a discussion.
I hate to show my ignorance, but why the focus on binary trees? I've been programming for decades, but there are plenty of things I've never touched because I just didn't have a need in the particular job. Binary trees are one of those things.
What I mean to say is, I have tended to use existing libraries that might make use of binary trees, but that I have never had to print or modify them. If modification is necessary, it's always through the library. So while I understand the basic concepts, I'm having trouble understanding why it would be a useful interview question.
Do developers at Google and Microsoft actually manipulate binary trees on a daily basis? Like, ALL of them? I can understand that they would be a common thing in the bowels of some code, but surely not all developers in either company would run into them on a regular basis?
Coding Interviews Exposed covers most questions you'll be answered.
It all depends on the team though, the interviewing process is completely non-regulated here. We get training telling us what to legally not say, but other than that is a free for all.
Teams ask white boarding questions because they heard other teams ask white boarding questions.
I recently went through an MS phone interview and didn't do as well as I would have hoped. I did learn quite a bit though. Here's my advice:
1) Actual programming and a technical interview are not and never will be the same thing. Assume you know nothing. And I do literally mean nothing.
2) Start with a basic intro to programming book in a language you aren't familiar with and try to do simple tasks in a language you already know. This simulates a bit how it feels to program under pressure. Things you know you know seem not to come out quite as quickly. This helps to emphasize point #3.
3) Practice. Practice. Practice. Practice. Practice. A decent chunk of your typical interview questions exist out there. I'm not saying to memorize them, but be familiar with their solutions. Eventually you'll get to the point that an approach for solving a new problem will come from pieces of old ones.
4) Since coding is done on whiteboards, do not use a compiler. Pencil and paper.
5) Find someone to give you an honest-to-goodness mock interview.
I'm currently interning at MS (India). I am a physics student with programming as a side hobby. (Not yet sure what to do as a career)
While this may be quite different from interviews for MS Redmond/etc full time jobs, this was my experience:
In my interviews, I mentioned off the bat that I wasn't great at algorithmic stuff. Now, don't get me wrong, I can do them, I'm just far, far behind the curve. (Students, esp. CS students, in India spend lots of time on sites like SPOJ and CodeChef solving algorithmic questions; I prefer to do stuff in open source). The interviewers were understanding. I got asked fun stuff like "what do you enjoy about [project I work on]". "Walk me through some large problems you grappled with on [project] and how you solved them". I got to have great back-and-forths on these things. It wasn't easy explaining nuanced and intricate problems, but I was able to give them a rough picture which I think is what they wanted. Shifting the focus from "can he solve this puzzle" to "what work has he done in the past" was great.
There were also algorithmic-y things, but not the usual kind. For example, one interview focused on designing and optimizing a build tool. The idea was that we had a bunch of components, with dependencies on other components, and I had to efficiently generate a viable order of building. There were some other constraints which I forgot. I realized that the problem was similar to a garbage collector (not that I knew how to model that either :P), and did some sort of mark-and-sweep thing. Then, after being asked by the interviewer, I optimized out some for loops and stuff.
I didn't need to know any algorithms beforehand. Sure, if I had known how `make` does dependency management, or had read up on garbage collectors, I would have done it faster. But I didn't need to know anything, and it worked out.
My fellow interns had really rigorous algorithmic interviews. They were okay with that because they were good at them. It's understandable that "solve this puzzle" is the default since not everyone has real world projects they've worked on (students). I don't like the status quo, but I don't think the interviewers will be intent on maintaining it either. If you have stuff you've worked on that you'd prefer to talk about, I think they'd like that.
If you think the interviewer will be open to it, I suggest mentioning it. They will know you're not from CS; you can mention that you don't have much practice with algorithmic stuff.
Major YMMV though. Also I've done very few programming interviews so the advice here could be totally misguided. :)
Even if you have used them like a few months ago. Very few people work on variation of the same problem over and over again, except in research.
Generally, you have a problem, you research, you experiment and then you wrap all that in a nice easy to use utility class or api and forget the details because there is something else you need to work on and that problem was just one piece of the puzzle anyway.
The worst hypocrisy in all that, is if you ask a specific question to a veteran developer on a average size codebase, he will very often need to look in the code and you will not think any less of him. But if it is an interview and he does not remember some trivia in a piece of code he didn't write - my god, he is a loser.
> How is it the best way if you are not going to be writing code on a whiteboard on a daily basis on the job you are going to be interviewing for?
It's not. I hate whiteboarding as much as you but I try not to think of it that way. This method is just a way to increase your chances of succeeding. I've not succeeded in every interview, for sure.
I do what I must to press on. :(
The best thing to do is not to take the rejection personally. I always ask, as nicely as possible, what areas I should focus on to improve for the future.
Sometimes I get no response but I often get replies, good replies.
It's about gaming the system to get a job. When you are in a position to hire other people, you can worry about a less awful method to select candidates.
> It's about gaming the system to get a job. When you are in a position to hire other people, you can worry about a less awful method to select candidates.
That's true. I have other features that reduces my chances of getting hired, but doesn't eliminate them. Still, I wish I had someone else's mind.
They all just seem to be varying levels of fanciness over relatively common data structures and algorithms. I was only applying for internships but the whiteboard stuff wasn't too bad when I did (and other questions I practiced online), I imagine almost completely because we'd done 2 related modules at university.
You spot patterns but there's so many different ways they can ask things and things that you just remember from studying it, that I imagine I'd do worse now unless it was close to what I've done since. On the spot I couldn't invert a BST right now.
I suspect you could probably pass a whiteboard interview a few weeks after taking a mooc on data structures/algorithms with enough depth for example.
I interviewed (unsuccessfully) at google. If you want to be prepped, yes, I would recommend going over the algorithm books, and make sure that you can do all the basics in your sleep.
That said, memorization, in my limited experience, won't be that helpful, because the questions will likely come with a twist that requires that you understand the algorithms and are good at adapting and applying them - and you need to be able to do it in 45 minutes on a whiteboard. It is very difficult.
I did find "cracking the coding interview" to be helpful prep, and really, the hard problems are fair game. Again, memorizing won't really help you, instead, view these as the kinds of problems you need to be able to solve in 45 min at the whiteboard.
I am kind of blown away that people can do this. I went home and did one of the questions I remembered from my interview with a compiler, and it took me several hours (though I didn't look up a solution, I did use the compiler constantly to check logic, and so forth, something that isn't available of course if you write on paper or a whiteboard). I suppose might be able to get this down to 45 min at a whiteboard, but it would take me a lot of time to get that sharp, months or more of study and practice.
None of this is mean to comment on whether this is or isn't a good interview process, it's just advice on how to prepare.
> That said, memorization, in my limited experience, won't be that helpful, because the questions will likely come with a twist that requires that you understand the algorithms and are good at adapting and applying them
Yeah...that is memorization. I've had questions like that before and the problem isn't adapting. Its the fact I don't have everything memorized. I have no problems taking an algorithm off a book/Google and implementing it in 45 minutes, even with some kind of twist.
The problem is I can't remember things I haven't used in 10 years so expecting me to do anything with them in an interview is kinda pointless. Sure, I'll remember the purpose of the algorithm, but I'm not going to remember the algorithm implementation details.
Yes, that is true. Memorization doesn't necessarily mean rote memorization - for me at least, the algorithms are too complex to memorize by rote, I have to understand them to "memorize them". But yeah, if you are able to implement certain data structures and algorithms "in your sleep", then this does imply, as you said, a kind of memorization.
I didn't get past the technical interview at google or amazon, so you might want to take advice from someone else, but here's what I think I'd need to pass the algorithm section of the technical entrance exam at google.
You must be able to do the following in about 5 minutes or less at a whiteboard with no errors:
1) print all permutations of a set
2) build and print a binary tree (preorder, postorder).
3) create a hashing function
4) quicksort and mergesort
(I'm leaving out linked lists, arrays, stacks and so forth because you won't be able to do 2 or 3 without them anyway).
The reason you need to memorize these (yes, I'll go ahead and use your word now, I agree with you that this is memorization) is that you can't be dealing with how to implement these if you're going to do enough in 45 minutes to pass the exam. I think that these three things form the foundation for most variants on recursion, sorting, traversal, combinatorics, and so forth.
Next, you need to be able to identify when something is a variant of these algorithms and solve it very quickly at the whiteboard. "Cracking the coding interview" is good practice for this, but in this case, I don't think "memorizing" these questions and answers would be especially helpful, because you're unlikely to get that exact question.
Here are a few of questions I've invented on the spot (not from my interview, not from cracking the coding interview) that you'd probably want to get close to solving in 45 minutes or so. In other words, if a buddy wanted to apply to a top tech company and asked me to give him some practice, I'd give him 45 minutes at a whiteboard for each of the following questions.
"In a binary tree with positive and negative integers, find and print all paths with a positive sum. Now assume the tree is ordered, and make your code more efficient."
"In an nxm matrix, find all sub square matrices with a positive determinant."
This might just be me, but I was kind of struck with just how much progress and accuracy you are expected to achieve at the whiteboard. And like I said, I am kind of blown away that people are able to do this. While I can pose those questions and develop a mental model for how to approach it, I just couldn't solve them at a whiteboard in 45 minutes to anything approaching what I believe is the standard.
Again, whether this is a good way to filter people out isn't something I'm discussing here, this is just my advice if you are planning to apply and want to be prepped. Also, there will be more than just data structures and algorithms on the exam.
> This might just be me, but I was kind of struck with just how much progress and accuracy you are expected to achieve at the whiteboard. And like I said, I am kind of blown away that people are able to do this. While I can pose those questions and develop a mental model for how to approach it, I just couldn't solve them at a whiteboard in 45 minutes to anything approaching what I believe is the standard.
That is why I stress its memorization because without memorizing things you've long since abstracted to libraries, you really can't perform in the stressful interview situation to the level the interviewer expects.
If, however, you take something the interviewee implemented in the past 6 months as the "standard problem" they could probably do it without much trouble.
Could you illustrate? I can find plenty of examples of mirroring/reversing a tree (below the root node) but none that really invert it. I haven't even found a definition of what inverting a tree means (aside from some academic papers I haven't been able to read yet), despite quite a bit of googling.
The fact is, that reliably separating bad from good engineers is incredibly difficult at best.
For big companies that want to scale hiring process, there's really not much you can do.
You put difficult questions and you hope that bad engineers will not be able to solve them. Then you absolutely have to commit to not hiring the guys who can't solve the assignment, otherwise it's all for naught.
I spent Xmas break reviewing the most recent TopCoder problems, C++ and STL, and read Sedgewick's algorithms book cover to cover. I drank two quad espressos right before the interview. The interview was a straightforward process and I didn't miss a single question. The offer came that evening.
I should have turned it down. I've written on this thread already what a career-churning waste of time my brief stint at Google turned out to be because I was blind allocated.
Fortunately, the skills that got me noticed by Google were in demand elsewhere then (and, ironically, in demand at Google now) so I just exited the building when it became clear HR had labelled me a troublemaker for trying to find a position that made use of my talents rather than opt for the career reboot they seemed to expect of me. Great perks, but the 2nd worst gig I've had my entire career.
A consulting company that hired me to develop PS2 tools and then yanked the rug out from under me on day one and told me I had to write a fast MPEG-2 codec from the ground up in the next 60 days despite zero zip nada null knowledge of MPEG-2.
I quit after a month or so. Later I learned their bait and switch tactics were S.O.P. and the company mostly existed to fund the CEO's various mistresses.
To all of those saying ranting on twitter is the wrong move I couldn't disagree more. Twitter is often the ONLY tool that the average person can use to communicate and/or call out large companies on their actions. This is BS and should be made known. Homebrew is an amazing tool and I'd be falling over myself getting the offer papers in this guy's hands if he came to me looking for a job. The fact that google turned him away only further cements my opinion that google would be a terrible place to work.
My main complaint with the tweet is that it's almost certainly speculation.
1) Most companies (for legal reasons) don't tell candidates why they weren't offered a job. Maybe it was because of the binary tree question, but maybe it was for some other reason.
2) Homebrew is a Mac-only product, so the likelihood that 90% of Googlers use homebrew is very low. Moreover, Google does not track the software its employees download onto their laptops, so there's no way they would even know the percentage.
> 1) Most companies (for legal reasons) don't tell candidates why they weren't offered a job. Maybe it was because of the binary tree question, but maybe it was for some other reason.
So when I was declined at Google I knew people who worked for Google. One looked up my profile in whatever system they used and the other simply asked the interviewer why. OP may have done the same but as you say it's unclear either way but knowing is still possible.
> 2) Homebrew is a Mac-only product, so the likelihood that 90% of Googlers use homebrew is very low. Moreover, Google does not track the software its employees download onto their laptops, so there's no way they would even know the percentage.
Most of the people I know at Google use Macs so I don't see why not. I don't know what the majority of them use but a good chunk at least do.
As far as tracking he could simply be tracking it himself and comparing his results of known Google IPs to employee counts. Not as accurate but might give a rough estimation. He may have further analytics based on data collected (if collected) by each machine.
I'm sceptical; I've heard this before but never managed to track down a source; I also don't know if you need some specific wording for this or if it's so vague as to be useful. Case in point: when I've failed my google interviews in London, I asked for the feedback, and got a reply saying something like "Sorry, we don't do detailed feedback. We weren't happy with your performance on the interviews". Which, duh, I got that from the being rejected part :P
Huh? I've been rejected by an UK company with just "you're obviously a very experienced programmer, but "we are not able to take your application any further at this time".
I did get a code review out of it though, and it did point out a few real issues, so I'm ok with it.
Maybe I don't understand the legals fully. My experience comes from being on the hiring side. Our HR department were very keen that we kept detailed notes so that the decision not to hire could be justified in the case of a tribunal or suing situation. For what it's worth, it did make me question my motives and felt that it ensured I gave every candidate a fair shot.
I don't have it installed on my Macs - what am I missing? I write C++ daily on them.
EDIT: Seriously, what am I missing that I'd need as a developer? My dayjob is on a Mac Pro writing desktop software and my hobby at night is on a MacBook Pro doing the same; it's on neither machine.
brew is a package manager, I can install a set of applications and libraries that will be kept up to date with a simple command line. If your coding doesn't take you outside of what Apple provide then you are probably not missing anything.
Personally I have the Android SDK and NDK, QtCreator, Qt5, cppcheck, cloc, mongodb, node, wine and others. If you are not using a variety of 3rd party tools to produce/automate your work then you probably don't need to worry.
Having all those kept up to date for me saves me so much time.
I use xcode and use wxWidgets, but I grab these manually and compile them myself. I also use the Android SDK and NDK, cppcheck cloc etc like you
I didn't realise that they could be kept up to date with homebrew?? Is that the case?
Last time I looked at MacPorts (and perhaps homebrew) I saw it as a porting of BDS/GNU tools to OSX and therefore making the OS more Linuxy to my mind. I wanted a clean break from Linuxland when I moved to OSX, hence I removed MacPorts etc. (Additionally the ported software all ran under X and not Quartz so it didn't really fit in; also, the fetched packages and compiled systems were massive in size which I didn't appreciate on a laptop that shared space with recorded audio and video etc. so space was at a premium).
Homebrew is really the default package manager. You don't need it, per se, but for the vast majority of people it'd be like trying to set up a Linux system without using apt-get or yum: an exercise in pain.
We don't have all the details. You might not hire the best computer programmer in the world if he was also a psychopath. Heck, you might skip on him even if he wasn't a good culture fit. I'm not saying this person is / was any of those things, but there's more to hiring than just talent.
There's also the committee factor. Google doesn't hire/no-hire based just on one person's opinion. If half your committee says, "90% of our engineers use his thing!" and the other half says, "But he can't code on a whiteboard!" then the committee is already kinda likely to take a pass due to lack of consensus. If there's any doubts at all on culture fit, which Google absolutely does care a great deal about, even more so. That said, when I was there, I interviewed someone who was essentially going to be my partner on something and everyone else (who weren't going to be working with this person) ranked the person poorly and I was the sole voice of (positive) dissent. I wasn't on the committee, so I don't know what their logic was, but the person was hired. And they were great. So it's not always even a majority opinion thing with the interviewers. In a nutshell, not only are we dealing with incomplete information here, we're dealing with VERY incomplete information.
That's not what happened at all. Google has an interview process that is designed to weed out false positives(incompetent people getting hired). Because this guy didn't have enough algorithmic knowledge he failed their filter. To work at the borg you must learn and play by the borgs rules!
Ranting on twitter is fucking awful. All it gives is nuance-free soundbites, and there are no 'lines to read between' due to it's brevity. Stuff gets taken the wrong way or out of context all the time.
One of the reasons why it's so powerful is because it gives lazy journalists a stream of easy reaction quotes. When else in the history of journalism has there ever been a private company that was so heavily promoted, regardless of the media source? And even then, the react quotes are poor quality, single-sentence shit.
This rant on twitter is absolutely meaningless unless you have a heap of context to go with it. And because twitter is popular, it's dumbing down public discourse along with it.
/rant, that I couldn't have fit in 140 characters.
maybe Homebrew is amazing, but Google can hire the best people in the world. that day they might've interviewed 10 people better than him who just didn't make a popular tool
"that day they might've interviewed 10 people better than him who just didn't make a popular tool"
Are you judging this on their ability to invert a binary tree? People here are ranting about having a problem with interview practices not representing whether or not you are good at your job. That's presumably (IMO) why the tweet was posted.
the tweet is made to trigger that discussion. but you can rephrase the tweet to something like: even though I made a popular tool, I didn't get a job at Google because I'm not good at other things they need.
Google probably has a lot of factors counted into the final decision, including algorithm skills, other software engineering skills, human skills, contributions to open source projects. It's never 1 reason why you get rejected, it's the total score.
Why do you think white boarding over the course of an several hours would be a better predictor of these traits than years of making very meaningful contributions to software?
It seems to me like creating and managing a repository of more than 4800 contributors is a much better signal of a candidates software engineering, communication, and human skills than a couple of people's opinion of a couple of hours of white boarding.
It's not that you're not _good_ at it, but you're not able to demonstrate under time constraints, lack of references, lack of iterative code/compiler/REPL feedback, and while being stared down by someone whose default mindset is "why am I wasting my time on this person?"
It's never 1 reason why you get rejected, it's the total score.
That depends on how far down the interview process you are. Often they'll cut you off at a phone interview if you stumble. Sometimes they'll cut in-person interviews short if they think you're not worth it. So, they don't always get a "total score" to judge against. It's often just summary judgement off subjective feelings.
It's also aggravating when a Google recruiter "reaches out" to you and but won't (or can't) reveal any details at all about what role they're trying to fill.
demonstrate under time constraints, lack of references, lack of iterative code/compiler/REPL feedback
well that is the interview. for you being good means to know how to invert a binary tree if you have access to a compiler and internet, but for Google it means to know it off the top of your head. they could let you have access to compiler and then make the questions much more difficult. which one is better?
I don't know. first one takes less time at least.
Often they'll cut you off at a phone interview if you stumble
I'd guess that's because they have so many condidates to pick from.
I'd guess that's because they have so many condidates to pick from.
Exactly. It's the Big City Dating Problem. There are so many people to pick from, nobody wants to compromise to build a relationship. Just leave at the slightest sign of something not being 100% what you want and on to the next one since there are 20,000 other acceptable people within walking distance.
if you have access to a compiler and internet
It's not exactly that, but you often don't realize how many micro-checks or micro-confirmations you do while programming. In the absence of any "is this okay so far?" feedback, you're left second guessing yourself and reading your solution redundantly to check for any errors your daily automation would catch immediately by default.
Disagree, this is not some PR campaign of a corporate drone, a real person felt frustrated with the process of hiring in Google. Is it to instigate debate on HN, I doubt it. He is just venting out, that his body of work which is public is effectively ignored in favor of a pedantic exercise.
I am speaking from personal experience here, they don't count a lot of factors while rejecting a candidate, you make one mistake on their particular set of questions, you are out. They don't even look at your resumes until all your 4 interviews are completed. However, selecting a candidate is a different matter, in that decision they might consider everything.
Does your personal experience include actually being involved in the hiring process at Google?
Because in my personal experience - as both an interviewer and a hiring committee member - what you said is completely wrong.
> they don't count a lot of factors while rejecting a candidate, you make one
> mistake on their particular set of questions, you are out.
Completely false. I have personally seen many people get hired despite doing poorly in an interview or making mistakes.
The hiring committee looks at all of the available information - resume, recommendations, and interview feedback when making a decision. There is no "one-thing" that will make or break a candidate.
> They don't even look at your resumes until all your 4 interviews are completed
I have no idea where you are getting this from. All interviewers receive the candidate's resume prior to the interview. I always review the resume's to get a sense of what the candidate has done so I can ask appropriate questions.
Yes, I was interviewed at Google last year. This was basically what happened. Sure, they have your resume, I am not sure if they read it throughly, the interviewer didn't even know my major until I told him at the end of the interview. While, they don't stop inteview and say you made mistake, it's competitive and it's not possible to hire you. But, you know at that moment that you are not getting hired. I guess, they need plenty of reasons to hire a candidate, but just one reason to reject. I am not complaining, it seems to work for Google very well.
And so Google can continue to lose talent to competitors, or have potential hires go off to start their own startups they can then pay a premium for, while keeping "good enough" talent for existing initiatives.
I'm starting to realize why Google is absolutely terrible at any sort of customer-facing product.
talent.. what is talent? they have a set of skills they look for. this guy doesn't have what they need. for you talent means to be the author of a popular tool, but for Google means more; means also boring stuff like being good at communicating, being good in a team, etc. who knows how many of these this guy failed
Esteemed schools of music probably produce dozens of guitarists "better" (in any way you would measure it academically) than Jimmy Page or Jimi Hendrix per year who go on to a life of producing nothing of lasting worth.
I'll take the guy/girl who has actually proved they can produce something useful every time over someone who is "better" but hasn't.
Well then I guess it's because Page/Hendrix were much more than expert guitarists. They were also good at composition, lyrics, showmanship, marketing, business. Those academic guitarists will make contributions to the guitar field but seems that you wouldn't value that alone and it's OK. You are free to have your opinion about whom to hire just as Google is. For the sake of the argument, you could imagine 100 reasons why Google wouldn't hire Jimi Hendrix. Even though he produced something very popular he lacks many things that Google wants: he's not respectful of other people so he's going to cause other people problems, he's doing hard drugs so he's unpredictable, good with guitar but not other instruments, and so on.
If Google keeps telling itself it only hires the very best guitarists, and it turns down someone like Hendrix - because he's "not a team player", or he doesn't know how to tap out 13/18 while improvising an Indian Harikhamboji raga, or for some other inane reason - then it isn't really interested in the best guitarists at all.
It's the difference between an holistic view of talent, and an academic and rather narcissistic insistence on specific limited social signals that are believed to correlate with intelligence.
And this really matters in practical ways. A lot of talented devs will be reading this story and wondering if they really want to work for Google. Negative PR like this is incredibly damaging. Consider the opposite - if Google could say "Yes, the Homebrew guy works for us." How much value do you think that would have had?
Long term, the really smart and inventive people start to stay away. And then you get something that seems to be happening to Google already - declining product quality, a poorer user experience, and diminished reputation. You know - Google's record of hits recently hasn't been that great?
So anyone who thinks Google is fine because it still has thousands of applicants is missing the point. Instead of being the kind of place where Hendrix plays, it's in danger of becoming the kind of place where there are a lot of people with a lot of solid but uncreative technical chops, and no one is making cool and interesting music any more.
I think this is spot on. I have seen 30 person teams composed of 30 one dimensional people, and 3 person teams composed of 10 dimensional people. With few communication barriers and everyone thinking about everything all the time, I take the small tight team. It does depend on the project, though, because a) sometimes the problems can be parallelized and many people working is good and b) 10 dimensional people don't typically like boring problems.
Has no one stopped to question what Google may have been looking for in a candidate?
The OP has written some great apps, sure, but there is a huge difference between writing a package manager for Mac (among other Mac/iOS apps and utilities) and writing incredibly complex, highly performant algorithms for say search indexing, machine learning, ai, etc...
In that context, knowing CompSci basics like Binary Trees (usually taught to pre-major CompSci students) has a lot of relevance. It's clear Google was looking for a Computer Scientist here, not just another developer.
I am also put off by the extreme display of non-professionalism here. It's like the OP took this personally as an insult. "I've written popular things, how dare you not hire me if I want to work for you!". That's not a professional I'd like to work on a team with. Rather, that's a throw-back to the "Rockstar" programmer that nobody wants to work with.
This boils down to a person who was personally offended a company did not hire them at a drop of a hat, as-if he was "blessing" Google with his presence.
I'd say, Google (and other companies) are better off without this type of ego.
They made the proper choice to weed this individual out.
I used to do interviews at Google, and most likely, the interview process just failed in this case. Typically a candidate will do 5 technical interviews, where the Google engineers can ask basically anything they want to (eg. invert a binary tree in this case) and score the candidate. If the candidate trips up on one of these interviews, that can be enough to reject them. The process will sometimes randomly reject qualified people.
Additionally, I think people have a mis-perception of the skills required to work at Google:
> writing incredibly complex, highly performant algorithms for say search indexing, machine learning, ai, etc...
...
> Google was looking for a Computer Scientist here, not just another developer
This is not really the case. The large majority of positions are for basically "just another developer" to maintain an internal Java tool, or some web UI, or whatever. They are complex in that they have many dependencies and often lots of legacy code, but not all to different from Microsoft, Apple, Facebook, or any company with a large code base. In my several years at Google, 0-5% of my time was spent coming up with complex algorithms.
Then why do they put people through those extreme paces? If you have the chops to get through the Google interview process, you are not "just another developer." If I were enough of a star to get through the process, got hired, and then was put to work maintaining some internal Java tool, I'd be pretty pissed off.
Yes, in fact that is exactly what happens. Many of the people at Google who are level 3/4 SWE's (about 70% of engineers are at that rank, it is considered a junior level) feel that they are under-utilized, and there is a lot of angst at that level. Those people would probably have more important/impactful roles at other companies. Some do in fact leave for other companies, but many end up tolerating the situation as other aspects of working at Google compensate (stable, brand name, perks, etc.).
As for why management continues this interview process: probably because they can. Lots of people apply to Google, and they have a high acceptance rate for offers extended. There is also a historical aspect to it: "we have hired this way in the past, and it worked, so why change it?" and fear: "if we lower our standards, the company culture is at risk" The problem of low-level SWE discontent showed up on management's radar around the time that I left, so I'm not sure if they are doing things to change that.
In fairness, package managers can get surprisingly complicated from a CS aspect. openSUSE wrote an elaborate reusable SAT solver library just for the dependency resolution component. Then let's not ignore that a package manager at its core must deal with the finicky nature of contemporary dynamic linking, the OS environment and all the baggage that comes with heterogenous library contexts and namespaces, maintaining transaction consistency, preferably ensuring atomic operations and so on.
That said, I don't think Homebrew is one of these. Always seemed like a quick fix for an OS X deficiency to me, and its competitors MacPorts and Fink being mediocre themselves certainly helped with its popularity. Not to rain on the achievements of the developers, but still.
To further your point, Package Managers have been done... and have been getting done for a long time.
Blazing the trail in machine learning, ai, etc... that's new territory where there are no knowledge bases or places you can go to source information about how to do it.
That's a different ballgame, and requires a scientist, not an engineer/developer.
Package managers have been done, but by and large not as well as they could be. It's unsurprising that Nix originates from an academic background. It's actually complicated work, and not really a solved problem yet.
Moreover, I'd heavily dispute that there are "no knowledge bases" for machine learning and AI. There very much is, given that the fields have a history as long as most contemporary computing. "Machine learning" in essence reflects the transition from AI being symbolic to statistical, so there's plenty of reference material on the latter and the various intersectionary disciplines like computational linguistics, data mining, computer vision and so forth. Hell, OpenCV - there's some good source information, for instance.
If you follow the twitter discussion, it says he applied for iOS tools. I don't know what tools they write but I'd be surprised if someone who manages to write a (pretty good, actually) package manager can't solve the problems in this position.
> If you follow the twitter discussion, it says he applied for iOS tools
We cannot just assume that since he applied for iOS something at Google that he's the best fit and what they were looking for.
Maybe the tools Google needs built are very CS heavy (hence the "Build us a Binary Tree" question)? Maybe during the interview his arrogant attitude was on full display? Maybe the interview team dug up past episodes of explosive arrogance like this very one we're discussing right now?
Even if they were looking for someone that matched his profile exactly, his post-interview display of non-professionalism is sure to hurt his chances of a re-interview anytime in the future (and quite possibly at many more companies than just Google).
He conveyed several things with this display, none good;
Google contacts plenty of people. Reaching out to yet another engineer (not even talking about this particular case) doesn't mean they really want someone particular. They just going through the pool of potential matches.
Person who initiated contact may not even know what exactly Homebrew is.
There is no slander in his tweet, he reported the FACT, may be a tad-bit dramatic, but FACT none the less. I just cannot understand why people think, Job seekers should cower down in public in fear not getting a future offer. To the hell with it, speak your mind, live like a boss even it means you make a little less financially, its better than being a rich coward.
- The guy was not offered a job at Google after his interview.
Anything else is unconfirmed and, speaking as someone who's thrown quite a few frustrated hyperboles onto the internet, sounds like frustrated hyperbole.
Frustrated hyperbole should not be taken at face value as fact; sometimes there's truth in a smaller version of what's said, but not always.
For example, there is no fact established that Google hired him because he can't "invert a binary tree". In actual fact, we don't know that Google even asked him to invert a binary tree (at least not specifically). It could be that they asked him a question that he thought was as irrelevant as academic datastructure exercises.
And we don't know that his answer was the reason he got turned down either. This is the part of the hiring process (and really any human interaction) that takes the maturity of recognizing that people and their motivations/reasoning are more complex than we reflexively flatten them out to be.
> There is no slander in his tweet, he reported the FACT, may be a tad-bit dramatic, but FACT none the less
What facts are you talking about? The "facts" coming from half the parties involved? The "facts" coming from an upset and irrationally thinking person who's caught up in the moment?
All that you say is true, but there's another side of the coin.
Nobody wants rockstars in teams, because rockstars are bad team players. That's the reason why they're rockstars.
Nonconformity (sometimes) leads to innovation.
Like it or not, a lot of the tools and libraries which we use every day and on which the internet runs, was written by lone-wolf 'rockstars' too weird and too eccentric to work in large corps.
You don't want your whole army to go over those mountains, you send scouts. Scouts are terrible soldiers, but they are quick to find a path between the trees and rocks and they come back with new information, new ideas and concepts of how to optimally win the battle.
Then and only then you bring the whole army and win.
A company, imho, should always have a small number of rockstars - impossible characters, crazies and lunatics with universe-sized egos, who have a track record of successful projects in the wild, working on insane things, alone or in very small groups, carefully managed by someone who can earn their respect - usually a lone wolf turned general.
They might be terrible programmers, they may have no education, but they have this unusual ability to see into the future and come back with interesting ideas and prototypes, which the 'soldiers' - people with solid fundamentals - can polish and transform into profitable products.
If a company innovates 'from the top', then it's trend is going to slowly be downwards towards irrelevance, because true innovation comes from the trenches and is done by the rockstars nobody wants to work with.
being able to invert a binary tree makes you qualified to develop machine learning, ai, etc code? that's a pretty low bar. If that's the level of knowledge you need to do that, then can it really be called AI?
If that's really what they intended for him to develop, then why not ask questions about machine learning or AI?
The OP is the type of person that, six months from now would be demanding special treatment because "he wrote Homebrew".
Writing a package manager (been done before, a lot of times) is far different from writing machine learning algorithms (new field, blazing the trail for the industry). One requires an engineer, one requires a scientist.
The OP flamed out because Google was looking for something he was not. Instead of taking it like a professional, he decided to retaliate.
The tone he took in his post is appalling, and I cannot believe HNers are defending him. "I'm special because I wrote something popular (and make absurd claims about 90% of people using it), so you better hire me or else" is what he just laid down.
That's not the type of person I'd want to hire or work with... and I'd wager most would agree.
Echoing what sseveran said, "You have done an awful lot of speculation here!"
Why do we call them "speculations"? Well, because those were your believes and most likely not the true depiction of who Max Howell is.
Have you worked or interacted with Max at all? I bet you not. If you did, you will see that he's not the egotistical and arrogant person that you depicted. [0]
To me, Max was just really frustrated and venting on the fact that Google passed on him because he failed in a whiteboard session of an algorithmic question, disregarding the fact that he has been proven to be a great developer who can ship code and a track record of teamwork and tremendous contribution to the open source community. And you know what, I am pretty sure Max can solve that whiteboard question with a little bit of preparation. It's unfortunate that Google interview process is designed to judge people on a simple algorithmic question. So, Google's loss on this one IMMO.
While I completely disagree with you on most of your judgements, I do agree with you on one point, that is, it's probably not going to be a fit (scientist vs. engineer). It's okay for companies to screen out experienced developers/engineers on academic knowledge, even though it's silly.
PS: I am not sure where the "90%" came from, but it's not "absurd" to believe considering Google's standard equipment issue has been predominantly Macs in recent years[1].
PPS: Max applied for an iOS developer role not a scientist role.
>Writing a package manager (been done before, a lot of times) is far different from writing machine learning algorithms (new field, blazing the trail for the industry). One requires an engineer, one requires a scientist.
I mean, a package manager is probably far more rooted in computer science than machine learning (which is basically just applied statistics: software engineering edition).
as someone who has worked on machine learning systems, this made me laugh out loud. not the fake thing you do when you mildly chuckle but don't actually make a sound. the thing where your coworkers look up and are annoyed. thanks!
I didn't agree with you at first.. but after thinking about it I think you're absolutely right.
Getting rejected from one interview at Google isn't the end of the world. The right and mature thing would have been to go home and figure out the problem that you got stuck on.. and then try again. We all go through this process, why should we expect anyone else to get a free pass?
I recently watched an online game of chess between a grandmaster and an anonymous individual.* The grandmaster got caught off guard and got lured into a trap. Instead of getting defensive, he sincerely thanked the other individual for a great game and spent the next few minutes after discussing the intricacies of the trap.
Shouldn't someone as accomplished as the creator of Homebrew feel the same way about a problem he got stumped on?
They reached out to him - he mentioned it in a tweet, and they've skipped the phone interview . Also, he would not be the one writing the machine learning, search indexing, but rather the app itself. All of those would be handled by other non-ios developers.
[Ex-Googler here] Truth be told this is a trivial question to be asked during an algo interview and as an interviewer I'd consider this a warm-up. Otherwise it's a rather poor question since either you know how to do it (ie, you have an idea about recursion) or you don't - there aren't too many shades of grey or possible follow-up questions that I can ask to probe the depth of your knowledge.
That being said if I asked this as a warm-up and we'd spend the whole interview trying to get that done then of course my verdict would be No Hire. As an interviewer it is not my job to look at your GitHub profile - instead I am assigned an area to check and I try to come up with the best understanding of the candidate in that area. While failing to reverse a binary tree is a total failure in algo/ds you can still be hired since there are several interviewers (if you make it to onsites, that is), each probing a different area.
My biggest problem with Google style interview process is that it's easily passed by folks who already passed it in a different company. After having interviewed hundreds of Google's candidates I moved to another big company with the same interview type and the experience was really surreal. On my algo/ds interview I got asked slight variations of the same questions I was asking myself - and over time I've seen some totally brilliant, unexpected solutions. Must have been a strange experience for the interviewer who got his questions each answered in 3 minutes tops. I also made it a sport to solve each one in a different language because why not. Not sure though about validity of the signal the company got from this interview.
> [Ex-Googler here] Truth be told this is a trivial question to be asked during an algo interview and as an interviewer I'd consider this a warm-up. Otherwise it's a rather poor question since either you know how to do it (ie, you have an idea about recursion) or you don't - there aren't too many shades of grey or possible follow-up questions that I can ask to probe the depth of your knowledge.
It is a terrible question.
First, you can't invert a binary tree (as in flip upside down). If you did, you'd end up with multiple roots and since all binary trees are rooted, you'd no longer have a binary tree. It'd be a tree, just not a binary tree.
If the questioner meant mirror a binary tree (swap left & right at each node), then it is a no-op. You do not need to modify memory to do it. Left & right are just labels for two pointers. You can just swap the order your variables in whatever struct defines your node and cast it against your existing memory (or change what your accessors return in a derived class or create a reverse iterator or however you want to implement it) and there you go, a "reversed" tree with zero overhead.
Either way, it is a terrible question unless you wanted to see if someone understood the difference between how a data structure is drawn on a whiteboard for humans vs. how it actually works. Maybe they were actually asking that question, but that seems highly unlikely.
And if they actually meant for you to recurse down and swap left and right on everything, it would dramatically lower my opinion of them because it would make me wonder if they knew the difference between how a binary tree is drawn on a whiteboard vs. how it is laid out in memory.
> First, you can't invert a binary tree (as in flip upside down). If you did, you'd end up with multiple roots and since all binary trees are rooted, you'd no longer have a binary tree. It'd be a tree, just not a binary tree.
Inversion is a transformation that maps directed graphs to directed graphs. Binary trees are a subset of directed graphs, so applying inversion to them is not unreasonable. That subset is not closed under inversion, so you can get results that are not binary trees, but I see nothing in the question that implies that the interviewer was asking for the output to be a binary tree.
That may even be one of the points the interviewer wants to see the candidate address. Since the output no longer has a single node from which all other nodes can be reached, in addition to just inverting it they may want the candidate to mention the need to have some new auxiliary data structure to keep track of the multiple root nodes (and perhaps note that in the inverted tree each node only has one outgoing link, so if we are inverting in place each has room for two, so we can use that now available second link space to make a linked list of the roots, so we don't need any extra storage for the new root list data structure).
> if they actually meant for you to recurse down and swap
While only the person asking this question knows for sure I'd assume that was the intention.
It's not a _terrible_ question - it's just a recursive equivalent of FizzBuzz. You would treat this as a way of making a (good) candidate gain some confidence. And that alone is extremely important since the more at ease the candidate is the better the signal that you're getting. What I was trying to say there was that if I asked that as a warm-up and didn't get a satisfactory response very very soon I'd realise that your man has no clue about recursion and would move to the main question, choosing one that does not involve it. It would be something easier than I'd ask otherwise and would be closer to real-life scenarios - like first solving a simple problem, then parallelising the solution and correctly managing concurrent access.
> First, you can't invert a binary tree (as in flip upside down). If you did, you'd end up with multiple roots and since all binary trees are rooted, you'd no longer have a binary tree. It'd be a tree, just not a binary tree
Just because someone wrote on Twitter with limited number of characters that the problem was to "invert a binary tree", does not mean there were no clarification from the recruiter.
> And if they actually meant for you to recurse down and swap left and right on everything, it would dramatically lower my opinion of them because it would make me wonder if they knew the difference between how a binary tree is drawn on a whiteboard vs. how it is laid out in memory.
If you're thinking about making a real structure, I believe you should recurse down and swap. If you don't do this, rotating, joining etc, could be a real pain.
If they said "swap left & right at each node" it would have been easy. But they used a term that doesn't even yield an example in google searches - what kind of "inversion" do they have in mind? It's as if everyone is on the same page but me.
"Invert a binary tree" and "inverting a binary tree" both give me as the first result in Google this Quora question where someone gives an example and asks how to do it: http://qr.ae/7NDCg4
At my university it's common to call acyclic, connected graph a tree. We distinguish between rooted and unrooted trees. For example, minimal spanning tree doesn't really have to be rooted, it just has no cycles.
Exactly. This style interview doesn't find the people I think are valuable - people who ship code. Instead, it validates people who spend time thinking about interview questions.
The link between the content of your GitHub and your ability to ship code is unclear at best. I knew a guy who had a really awesome GitHub profile but turned out to be a mediocre contributor to the codebase. He claimed that coding style guides we used were hurting his creativity to the point that he did not enjoy coding and would avoid it if possible. He'd fix bugs and make small-ish contributions but would never ship major features or libraries "because of the style guide". While I understand the sentiment I always found his attitude juvenile.
The thread is about the industry in general, which has adopted Google's algorithm heavy interview process, but also adopted github and open source as a resume.
WTF is "inverting a binary tree?". The smattering of search engine results points to some seemingly operation that basically generates garbage by destructively manipulating the tree into a DAG in which the leaves of the original tree are roots, which point downward the parents. The original root node is returned, and still points to its children. Unless you return some aggregate (e.g. list) of all the roots which are not direct children of the original root, or it is understood that something elsewhere holds reference to them, they are leaked.
Be that as it may, if the interviewers hand you a reasonably detailed specification of what "invert a binary tree" means and you can't whiteboard it, I don't see how you can expect the job.
If you're expected to know what it means, and get no hint, then that is just stupid. "Whiteboard up an implementation for something we refuse to specify, beyond citing its name."
I agree. I have no idea what "inverting a binary tree" means, or why I would want to do it, so I'd need some context from the interviewer. Chances are, I'd ask some very pointed questions about whether a binary tree is really the best underlying data structure for whatever the "inversion" process is intended to accomplish. The interview would probably go downhill from there.
My guess is that the potential leaking of some nodes is one of the points of the question. A lot of people would not notice it.
Note that in the original tree each node needs storage for two pointers. After inversion only one is needed. You can use that now unused storage to link together the multiple roots to solve the leak problem.
But note that only consumes the second pointer storage in the roots. Interior nodes end up with some free storage after inversion.
Since inversion is reversible, the binary tree and its inverse contain the same information. Does this mean we should only need one pointer's worth of storage in the binary tree nodes?
Is there something for binary trees similar to Knuth's xor trick for doubly linked lists?
If you invert the tree such that each node points to its parent, and there are no child pointers, you lose information. Namely, the order among children is lost. A given interior node still has two children as before, but it is not known which is the left child and which is the right child; there is just a set of up to two nodes which point to the same parent.
The binary tree inverse specifications/implementations I have looked at preserve the information by selectively using the left or right links. For instance, given:
P
/ \
L R
The inverse would be:
L R
\ /
P
Both children point to the parent. But the left child uses its right pointer, and the right child uses the left pointer. That's what preserves the information about which child is which.
Sorta ironic, but remember that the point of an interview is to determine how well you'd do in the environment that's hiring you, not how good a developer you are. Because Google tends to reinvent the wheel for basically everything, algorithmic knowledge really does matter. Package management only matters within a few very specific subteams. If you're interviewing for a general SWE position, you could be Mark Zuckerburg and still not qualify for it.
FWIW, I find Facebook turning down Jan Koum in 2009 and then spending $19B to acquire his company even more ironic.
That may well be, but if you're considering hiring a guy who's an expert at writing package management tools, don't you put him on one of the specific subteams that needs that skillset? Surely it's something Google deals with?
The problem is that Homebrew is very different from the sort of package management tasks that Google deals with. The design goals for Homebrew include: make it easy for users, make it not require root, make it work on Macs, handle dependencies robustly. If he were at Google it would be Linux-only, he'd be using Linux containerization extensively, he'd be deploying packages to thousands of machines instead of one, it'd be a virtual guarantee that some of the machines would fail during the installation process, there would be little or no user intervention and if there was a user it'd be a trained SRE, and the installation procedure would probably need to be an order of magnitude more efficient than Homebrew is.
I don't want to take away from his accomplishments as a programmer - I use Homebrew too. But my point is that it's very easy to see "Good programmer, of course he should get hired" from the outside, while the reality is that it may not be all that similar to the tasks he'd be doing.
Last I knew, Google corporate Macs don't include Homebrew. (I left Google a year ago, things could've changed since then.) I'd be very surprised if it was allowed, I have an open-source project that's available via brew and I've had people at other companies tell me that anything brew-only is a complete non-starter at their company because Homebrew isn't allowed for security reasons, and Google cares a good deal about security.
I assume that the dev was referring to Googlers' personal laptops with the "90% of your devs use Homebrew". It's not unreasonable to estimate that 90% of Google engineers use a Mac as their personal machine, although 90% using Homebrew is probably an exaggeration. A lot of Googlers use their corp machine as their personal one (they're not supposed to, but it's hard to police when all you do is surf the web), and the majority of them don't have side projects outside of work. I didn't install Homebrew until after I left the company, since before then all my code belonged to Google.
> If you are hiring the Homebrew dev, and your devs currently use Homebrew, why wouldn't you hire him to work on Homebrew for you?
Macs account for something like 8% of the total marketplace of all PC's. For developers, they account for something like 20%... another 20% on Linux, and remainder on Windows or other.
So even if somehow having a paid Google employee work on Homebrew seemed advantageous, it would only benefit 20% of Google's staff, and 0% of the company itself (all Google servers are Linux).
It might except ex-googlers in this thread of stated Google "banned" use of Homebrew internally. So the net benefit to the company and/or employees remains small to zero.
This is off topic though, since he was not being interviewed for homebrew development.
As far as I'm aware, Mac is just as 'banned' as Windows. Developers have Linux(Goobuntu) desktops and Mac, Linux, Windows or ChromeOS for laptops (which are primarily used as a terminal for your desktop).
I failed my third google interview. My problem wasn't that I couldn't solve the problems. I just wasn't fast enough because I didn't spend 3 weeks beforehand training speed in programming brainteasers.
All of infrastructure. You end up using LSM trees, Paxos, variations on binary search trees, checksums, Reed-Solomon, bloom filters, external sorts, and all sorts of other academic algorithms when you deal with distributed systems.
Docs/Drive does a lot with operational transforms and CRDTs, ever since Wave was merged into it (c. 2011). Remember that all Google Docs are collaborative.
All logs analysis & experimentation relies heavily on statistics, as well as online algorithms for reservoir sampling, quantiles, and a number of other statistical quantities.
Machine-learning is pervasive, throughout Shopping & News (it's all over these), GMail (spam, priority inbox, categories), Search, Ads, self-driving cars, Brain, and many other areas of the company.
Chrome and Android both do a lot of traditional OS & systems programming.
The main product area that didn't use algos heavily was Google+, although even this had some interesting stuff in stream ranking and search.
This is one thing people don't understand. In my current work I need to regularly design a new algorithm or modify existing ones. I've to keep books like CLRS and TACP always accessible, not for show but actually looking through pseudocode and analysis. I regularly sweep through dozens of research papers to find the state of art. These are not one-off events it's my and my teams day to day life. Lot of people often gets surprised when I tell this to them because most developers think they never need to use any CS stuff they had learned. If they ever need answer on any algo there is always framework or wikipedia. They just don't understand that most real world problems (or at least the ones we work on) require significant amount of customizations and mixing. These same people then insist that just because they have "shipped" some code or because they have showed "coding" capability via github profile, they should be eligible to be hired anywhere. Coding and development is "given" and that's considered as easiest part in our line of work. Shipping things is certainly art but that's hardly sufficient condition. The choice, analysis and design of algorithm makes ALL the difference in type of problems that we work on. If a guy showed up who wrote some package manager on Mac but couldn't work with binary trees, I would probably would pass too because in the kind of work we do his chances of success is fairly low. Frameworks and programming languages can be learned pretty fast but learning CS and building problem solving abilities using CS primitives takes years and years of intensive study.
There's totally a lot of crunchy CS in many things Google is into (and I do distributed systems stuff, so I can empathize heavily with your first paragraph). If you're doing it right, though, writing the crunchy CS bits of these should be exceedingly rare. Understanding is important, but I'm not sure you get a measure of working applied capability from writing this stuff in an interview. Discussing it, surely, but this is stuff that one--wait for it--Googles when dealing with implementations. And I strongly doubt that the overwhelming majority of Google developers are involved with the actual construction of this stuff.
The real reason, as near as I can tell, is because it allows management to not think as hard about allocating talent (the mythical "fungible engineer" nonsense), but whether it even works is another thing entirely.
Not everyone is constructing them, but everyone needs to be able to reason about they behave, determine what they need, and yes write it if a good option isn't available.
I'd like to present HN with a challenge. Come up with an interview process that matches _all_ of these requirements:
Objective - the process avoids human bias (this guy was in my frat, therefore is awesome).
Inclusive - someone doesn't need extensive past knowledge in a particular area of coding to do well in the interview.
Risk Averse - Avoids false positives.
Relevant - it properly tests the abilities needed for a coding job, and not irrelevant ones (like the ability to write an impressive resume).
Scales - this will be used for tens of thousands of interviews.
Easy-to-do - this will need to be done by hundreds/thousands of engineers, who would probably rather be coding.
It's easy to poke fun at what is perceived to be a flawed process. It's much harder to propose a solution that satisfies the above requirements. Google has done extensive research on this topic and has done remarkably well with it compared to other companies of similar size.
Xoogler here. I can't meet your challenge. IMO the very premise of a company-wide unified interview process for all software engineers is wrongheaded.
How can you make the interview relevant without knowing what the position is? I was asked the typical CS-type questions in my interview, but the team I ended up on required no theory.
How do you define false positives before you know what the candidate will work on? A superstar in one team will be a dud in others.
And let me add another bullet point to your process wish-list: gives the candidate a sense of whether they want the job. This is impossible when the interviewer is a random engineer from an unrelated team, unable to speak to what the candidate's work life will be like. A Google style process gives candidate very little information.
I would instead propose something very old-fashioned: teams hire for themselves. The usual reply is that this results in an "inconsistent hiring bar", but so what? Teams have different requirements and need engineers with different skills, so why shouldn't the hiring process reflect that? We are not fungible engineering units.
Create an online game/project/test that people must solve in order to get an onsite interview. The game can be customized for individual teams to have more appropriate domain challenges. Spend the onsite interview walking through the code the people wrote for that game. Devil is in the details, but that's the broad strokes of what you are asking for.
It may be hard to find a solution to the full interview problem. I still do not prefer whiteboard interviews; would be happy to code using my own laptop, projecting the session for everyone else in the room to see.
To be fair, inverting a binary tree is a pretty easy question. Google also tells you BEFORE you start the interview process that it'll be very data structure/algorithm oriented and asks that you please prepare (and take as much time as you want doing so). They even say that they want you to prepare because they know a bad candidate that prepares can look better than a good candidate that doesn't prepare - then want all candidates on a level playing field so they can make accurate judgements. All that being said, I still think that there is lots of room for improvement in the process.
I am dismayed by the way all the reactions on Twitter are piling on with outrage and/or relating similar experiences.
Inverting a binary tree is pretty easy. It is not quite as trivial as FizzBuzz, but it is something any programmer should be able to do. If you can't do it, you probably don't understand recursion, which is a very basic programming concept.
This isn't one of those much-maligned trick interview questions. This is exactly the kind of problem one may have to solve when writing real software, and though you may never have to do this specific thing, it is very related to a lot of other things you might do.
I run a small software company and I very likely would not hire a programmer who was not able to step through this problem and express a pseudocode solution on a whiteboard.
> If you can't do it, you probably don't understand recursion
No, I can't do it (don't even understand the question) but I certainly understand what recursion is, and can solve problems and make things work far more reliably than many of the more academic programmers I have worked with.
Well one of the points of an interview is to see how (or if) the person goes about understanding problems the context of a given problem. So presumably it's unclear to inverse a binary tree to you - that's why you ask clarifying questions. Perhaps you draw what you think it means and ask if that's correct. What will be considered the root if it's just a matter of reversing the direction of the pointers? Stuff like that. Google isn't dumb, and I'd be willing to bet that it's questions like those that are just as important than coding the solution (since it won't be that hard once you clarify the problem).
Here's an example to think about: You're asked to write a method that, given a word and a book, will return the number of times that word appears in the book. A logical way to continue would be to (after clarifying what a book is represented as and things like that) write a simple method that iterates over the book and counts how many occurrences of the words there are, and returns that. Is that a good answer? No, because you failed to ask questions like whether the method might be called many times for the same book (in which case it'd be better to build a map of words to occurrences). Or maybe the interviewer would say that it'd only be called once per book, in which your solution is perfect. The bottom line is that coding isn't as important as the way you approach the problem to the interviewer.
The best analogy I've heard for this: If you wanted to judge whether a person was a good chess or card player by seeing just a couple moves or hands - you wouldn't just look at the couple moves/hands played - it'd give you very little information. But if you asked the player to explain their whole thought process for the moves/hands, you'd get a much better indication of how good or mediocre that player was.
I'm a CS major, and I'm not even sure exactly what they're expecting. I mean a binary tree is just nodes with no more than 2 children. A linked list is a binary tree. Do they want a search tree? Do they want a self balancing tree?
Can you confirm that actually inverting a BT or solving other similarly probs would have been part of his future job at Google?
If so, they were consistent in disqualifying him in spite of his awesome track record, fame or "street cred" but if not, I can't really justify or defend this type of ceremonial or bureaucratic gotcha interview questions as they are just useless and can be abused by the interviewer to reject applicants they dislike for any reason or whim they might have at that moment.
But my point is this is not a bureaucratic gotcha question. If you can't do this task, you don't really know how to program well. Sorry but that's just how it is. It's like failing FizzBuzz.
There is this culture of crappy software that has happened lately, especially in the Web world, and it is really quite lamentable. I believe that a very large positive impact would be made on the world -- due to the extreme prevalence of software these days -- if more people would take seriously the idea of software creation as a craft with a very high skill ceiling, and work diligently to improve their understanding and their skills.
Note that "invert" in this context is fairly ambiguous. I suspect the interviewer meant reverse, which is, I agree, utterly trivial.
I, and apparently many others in this thread, spent some time trying to ferret out an answer to what it would mean to invert a binary tree: at first thought it would imply a collection of nodes all pointing at their parents, which seems not very useful (and would require more than 10 lines of code to whiteboard reasonably).
Much of the discussion is about whether or not the problem was described in sufficient detail or not.
The question is definitely not described in sufficient detail. That's usually on purpose. The interviewer WANTS the candidate to realize that too and ask the questions necessary to actually understand what they are trying to solve. It's very analogous to nailing down specifications on a new feature or something like that.
I've never inverted a binary tree (I'm not familiar with the concept). You claim that makes me a bad programmer. Be careful there: just because somebody hasn't been exposed to something doesn't mean they aren't smart enough to figure it out. Fizz buzz is also flawed in this way: it doesn't test if you know how to program, it merely tests "have you been exposed to the modulo operator yet".
FizzBuzz is not about the modulo operator. You can write it pretty easily without it, in several different ways. If you can't figure out how pretty quickly, you're not a good programmer. Sorry.
It is even kind of reasonable to just substitute a call to some black-box is_divisible_by() if you can't figure out how to do that test ...
No one is claiming that makes you a bad programmer. However, if you got this as a problem to solve and your response is "I don't know" and you stop there, then you are a bad programmer. As I've commented elsewhere, the interviewers want you to ask clarifying questions. Those are often just as important as the code itself. Once you've nailed down the definition of inverting a binary tree (maybe it's just switching left and right node pointers), coding will be trivial. It's much like getting sufficient detail on a feature request before starting to work on the feature.
What does it mean to invert a binary tree? I'm not familiar with this operation on binary trees. Does it mean to swap parents with child nodes? Or to swap siblings?
I have a feeling that the question is not really about whiteboard coding or solving graph problems. This is a blatantly ambiguous task. Maybe it was meant to force the interviewee to demonstrate how they communicate to the interviewer that they needed more information about the problem before starting.
Pretending you understand or blindly assuming what the interviewer meant is more of a red flag to me than requiring a few rounds of clarification.
In my opinion, an ideal interviewee would communicate that they don't fully understand the requirements and possibly offer a brief set of possible interpretations. Once the interviewer has clarified, my ideal interviewee would restate the clarification to demonstrate competence and verify their new interpretation of the requirements before proceeding.
In this case, it means to reverse the binary tree, so that you get the largest item by iterating down the left branch to the bottom of the tree. You would do this by breadth-first scanning the tree, swapping the left and right pointers as you go.
Oddly enough, Google shows me interview questions where "inverting a binary tree" means something quite different — for example, flipping it upside down, and making the left leaf the root, the right leaf the left leaf and the root the right leaf.
If this was really about "reversing" the tree, as you mention, the question seems more likely to address how the candidate approaches the situation. Like, he should start by making sure they both agree on what the question actually means.
Once that's out of the way, it seems relatively easy to come up with a naive solution, without having memorised any algorithms. It seems more like a case of brainfreeze to me, which can be sort-of fixed with practice (which in turn many candidates refuse to do: the dreaded "If I have to cram for the interview, I don't want the job" statement.)
So maybe he really wasn't a good fit for Google, despite apparently being a rockstar developer. Hey, startups need rockstar devs too.
I was gonna say, who calls reversing a tree's ordering inverting?
> Oddly enough, Google shows me interview questions where "inverting a binary tree" means something quite different — for example, flipping it upside down, and making the left leaf the root, the right leaf the left leaf and the root the right leaf.
Whaaaaa? I can't find this, but it seems like such a weird operation. Got a link?
I don't even understand why you would ever want to do this. You could just invert the input or output at the beginning or end (depending on what the binary tree is used for) for super cheap. It's not uprising that no one has done this in practice, though it also seems trivial (visit each node once).
It's an interview programming question, so what matters is that it's a good interview programming question, not that it's something you would ever want to do. It's an important distinction. Generally you're looking to test candidates' abilities to come up with efficient algorithms to solve problems, which you're not going to get if you ask them to, say, tweak the order that fields are displayed on a web app form (even though that's a "real programming task").
Although I do see this as something you could possibly do "for real". Sometimes you'd rather change your existing data once than continue carrying forward what amounts to hacks, which is what inverting the output would amount to. Reversing a binary tree would be what you'd run in a tool to migrate all of your data over from the old format to the new format.
Yeah I checked Google as well, but swapping children seemed too easy for an interview question. I'm wondering if it means restructuring the tree so a different node is the root. Trees have the property that any node can be the root and it is still a tree. Not sure just wondering if anyone had more info.
It does seem way too easy. Maybe it was the first part in an increasingly difficult several part question, each part building off of the previous part, and the interviewee never made it past the first part in the time allotted?
Interviewing seems broken to me too. I have spent the last few months trying to get a job out of college and everyone seems to be interested in how well you can regurgitate CS fundamentals.
They are seemingly less interested in seeing how you solve problems and work through the process of software.
I would be more than happy to see this process change, I am just not sure what it entails.
Devil's avocado: if they're truly CS fundamentals, then they should be baked into you good and deep during the course of your college education. It shouldn't be painful at all.
Maybe you're just not doing serious programming. Most people I know implement data structure searches quite often.
If you're writing scripts, or JS code for web pages or something like that, then maybe you don't use CS stuff, but ... are you able to write a web browser if you had to? Are you able to write an operating system or navigational software for a spacecraft? If not, then maybe just see this as revealing sectors of your skill set that could be beefed up, rather than presuming that none of that stuff is important.
> Maybe you're just not doing serious programming. Most people I know implement data structure searches quite often.
Wow. Really? Most serious people I know use other people's implementations that have already been highly optimized and well tested because they have better shit to do than reinvent the wheel.
I suppose if you want to write your own red-black tree from scratch, that's your prerogative. The last time I did was 20 years ago and not only will I never do that again, I will laugh at anyone who does it without a damn good reason.
> Wow. Really? Most serious people I know use other people's implementations that have already been highly optimized and well tested because they have better shit to do than reinvent the wheel.
Ditto. Those who decided to reinvent even basic data structure stuff left me a huge string of bugs to fix, which I eventually got so fed up with, that I started replacing their code wholesale with off the shelf solutions to stem the flow at my last job.
Aside from fixing an untold number of implementation bugs, the replacement caught several usage bugs as well, due to actually having some error checking built in.
We had just plain broken hashtables, "lock free queues" that didn't use memory barriers... or interlocked intrinsics... or even volatile, if my memory is correct - and not a debug visualizer to be seen before I got my hands on them, of course.
> I suppose if you want to write your own red-black tree from scratch, that's your prerogative. The last time I did was 20 years ago and not only will I never do that again, I will laugh at anyone who does it without a damn good reason.
Besides laughing, I'll tend to -1 the code review as well.
In many performance-programming situations you are subject to constraints that prevent use of a general solution or else makes use of a general solution massively inefficient.
For example, your data structure is on the GPU and your data is in a texture in a certain specific format because of other reasons.
If you wrote the above reply without considering this kind of case, it probably means you haven't been exposed to very much of this kind of case ... ... which was my original point.
The question isn't whether you should reinvent the wheel, it's whether you can if need be. It is important for companies to weed out candidates that can talk a good game but are useless when it comes to actual programming. So maybe in the process good candidates fall through the cracks, but then again, nobody has invented a perfect way to interview candidates.
In any case, how would you even go about knowing whether a certain piece of existing code that you blindly pulled in from some random place is optimized for your particular use case without first being able to conceptually understand it - which is all an interview is. Nobody expects candidates to create production ready code on a whiteboard.
Nobody's advocating rewriting your own red-black tree from scratch, for no reason. People applying data structures talent and skill aren't copying down algorithms they have in the standard library.
CS fundamentals like BSTs are definitely not used to teach programming--things like video game projects are--they're used to teach algorithm design and analysis. CS is an academic science, it's not a technology trade.
I think a more fundamental skill these days is figuring things out. For instance, a test where you actually have to google for the docs to something and figure out how to use it might make more sense.
The idea that "serious programming" means abstract data structures or coding in assembly is weird to me. I would consider a lot of people serious programmers who, while they may know how to do those things, don't actually do them almost ever.
What material did you spend time on that you don't know how to reverse a binary tree?
It isn't a difficult problem, even having never seen it before. This is one of those warmup problems to test how comfortable a candidate is with basic concepts such as recursion.
Especially as a fresh grad. Reviewing for a week or two if you've gone to a good school/actually did your work instead of 'consulting google' is definitely enough...
If you are trying to get a job straight out of college and you didn't have many relevant internships, there really isn't much to judge you on other than what you learned in college. This isn't a fault with the interview process.
Some companies give take-home assignments (small projects that can gauge how well you learn). I tend to find that these are smaller companies. You can always try there.
Also, the most important thing that they don't teach you in school is that getting a job is all about connections. Email some people that you've interacted with to see if they know of anything that might be suited for you.
I guess it depends on where you're interviewing and what their expectations are of you. Since you've just graduated they may expect you to be very junior and just want to make sure you have a good CS foundation. They should ask problem solving questions but it wouldn't be weird if they just focused on what they think you know.
I've phone interviewed with Google a couple times. I wasn't really interested in working there, but wanted to see what it was like. Both times, the people who interviewed me were decent, friendly folks and we had a good chat. They then dug into algorithm questions on topics I hadn't seen since my undergrad (I'm about 20 years into my career) and haven't touched since then -- though I've done a bit of algorithm work outside of that they weren't particularly interested in that.
I reached into my way-back machine and tried to derive some approaches where I simply didn't remember the answer (and I was very open about it). I made it to call backs both times, but they declined to move forward, probably because they wanted a younger person who remembered their big-Os off the top of their head, but I was okay with it. I told all the interviewers I had fun and I did.
Even if I had made it in, I'm not sure I would have taken the job at those times. So my lack of motivation ended up turning what could have been stressful into a fun look at their hiring practices.
However, I can see for people who are really dead set on working there, it can be a harrowing experience.
If you rate engineers on their ability to memorize something they could have looked up on Google in 5 seconds, what are you really actually testing?
If you're not at Google's scale, a take-home test that involves them producing a working solution to a problem in their own time, that you then expect them to be able to discuss, reason about, justify, and so on is a much better tool; even at Google's scale, where you might be worried people who copy answers off the net, have them bring the laptop they used in, and then get them to expand upon their solution.
All that said, I'd be surprised if that was the only factor at play. The Twitter comment itself seems to indicate a - uh - cultural fit issue. Additionally, and with no actual knowledge of the insides of HomeBrew, that a piece of software is widely used isn't evidence that the author is an excellent (or even good) developer.
Rambling weirdness stream of consciousness, but, yeah you might well be right... how hard can it be to reverse a binary tree? Isn't the "real" solution to sound out the solution as you go, and let the interviewer know what you're thinking?
I've never attempted to do this before, and my compsci is real weak, to the point where I'm going to make an educated guess about what a binary tree even _IS_ so... Let's start with what is a binary tree? Presumably we have a tree where every node has zero or two child nodes. Clue is in the name. In fact, that's an implementation detail, so let's say we have a tree where every node has a place two hold a left and a right node, where either can be null. Guessing that in order to be useful, the nodes need to hold a value. We need to choose a language here, and I know a miniscule amount of Go, and surely the cool kids at Google like Go, so:
type Node struct {
Value int
Left *Node
Right *Node
}
Now what does it mean to "reverse" a binary tree? Are we simply swapping all Left values for Right values? It feels a bit like there might be a trick here. So let's draw a binary tree with some numbers, and think about what reversing it might look like:
Actually, that looks pretty reasonable to me, and that's a straight flip of each Node's pieces. I had wondered if I might need to only swap at every other depth, but this just looks like simply swapping over the left and the right branch will work fine.
We can either iterate or recurse here, right, because it's a nested data structure, and I once read almost the first two chapters of SICP, and I'm pretty sure those are the options. So, let's have a quick think about the memory / performance implications of that. I cannot think of a way we can avoid visiting each node here; perhaps there is one, but cards on the table, I can't think of one. So let's assume we have a runtime cost that will grow linearly with the number of nodes. If that's the case, I'm pretty sure that's what the cool kids call O(n) - maybe that's O(1), but I'm sure Wikipedia can tell me.
So the mechanics of switching any given Node looks pretty easy:
nr := n.Left
n.Left = n.Right
n.Right = nr
That should handle nil no problem too. But we also have to switch the children. If my original assumption about having to talk to every node is correct, what the question really wants to do is know if we can do this in a way that isn't too much of a memory hog. Now a simple recursive method would do this, for sure:
func (n *Node) recursiveReverse() {
nr := n.Left
n.Left = n.Right
n.Right = nr
if n.Left != nil {
n.Left.recursiveReverse()
}
if n.Right != nil {
n.Right.recursiveReverse()
}
}
But while I know _NOTHING_ about how Go's callstack works, I bet it doesn't do magical optimization to stop it simply growing and growing and growing, which I think the cool kids call "tail optimization". At this point, one turns to the interviewer and says "How big do these trees get?", and makes some point about simple and elegant solutions to problems that make developers lives easy, because hardware and processing power is cheap. Let's assume the interviewer doesn't let you off the hook so easily...
Let's take a second to think about the memory consumption of what we've got here. At the moment, we add a new ... thingy (are they called frames? I have no clue) to our callstack for each child, and our worst case scenario then is the maximum depth of the tree. No-one has told us if our tree balances, but I suspect that if it does, you could then describe the max-depth using a formula involving `log` and the total known nodes. Wikipedia would know.
What else could we do? You could keep an explicit stack of nodes you knew you still needed to be visited, and go depth first. My gut feeling there is that you could write this where you get a worst-case scenario of your stack getting as big as the deepest node, plus one or two. If you wanted a truly iterative approach, you would need to store a pointer to the parent in each Node; there's a memory (or storage) cost to that too, so how often do we actually need to reverse the binary tree? That's a question to think about.
Given all this, then, one comes down to: what's the relative memory pressure that each frame in the call stack exerts, as opposed to pushing a pointer on to the end of an array in Go? Oooh, Go arrays are immutable, so actually you need to think about slices, unless you already know the size of the tree, and can preallocate exactly as much memory as you need. I wonder if there are memory implications there? With my commercial hat on, again, how much does this actually need optimizing?
Anyway, those are the considerations that come to mind; it would be worth checking Wikipedia for a nice iterative approach here, or if there's a sensible known algorithm. I'm going to stick with the recursive version I had above as my whiteboard version.
It can't be just mirroring, because there's the obvious zero-op solution because "left" and "right" don't actually mean anything except when you're visualizing it for humans:
This is genius. My C is rusty, but isn't this more than visualising for humans, if I cast a NormalNode as ReversedNode, traversal should be reversed, right? (I.e.reversed->right is normal->left
Yes, everything would be reversed. That version isn't actually how you'd do it if you actually for some reason wanted to use it in real software, it just was for illustration purposes to show "left" and "right" are just names and you don't need to change anything in memory to flip them.
A similar effect could be done in an OO language by having a base class NormalNode with two accessors, getLeft() and getRight() work normally and then a derived ReverseNode where getLeft() called super.getRight() and getRight() called super.getLeft(). With a proper compiler, the whole thing would simply be optimized away.
>If you rate engineers on their ability to memorize something they could have looked up on Google in 5 seconds, what are you really actually testing?
Repeating what I said earlier - You can't build a plane by opening a physics textbook. Starting with a LARGE pool of fundamental and higher order building blocks allows you to fit them together using whatever skill/talent you have at solving problems, in a MUCH shorter time than someone who doesn't.
Someone who constantly google's basic stuff is a net drag on a team's productivity. Rather than operating at their skill level, they're constantly being distracted by having to task-switch and waste hours researching stuff. The internet isn't going to magically know the exact problem you're working on and whether certain data structures or algorithms are appropriate for solving it, or if they need further tuning depending on what your priorities are, or what have you. Getting questions answered, and googling/learning about CS stuff is great on its own, and should be encouraged as much as possible, but it is not a substitute for actually retaining that knowledge.
Interviews break most people's intellectual context. Talky/show-me interviews are designed around interviewing MBA "how well can you present yourself" roles, but it completely fails to evaluate a technical person.
Doing a blind and "surprise me with a clever (or traditional) algorithm" whiteboard interview is a teaching skill, not a coding competency. You have to illustrate, explain, make your self vulnerable while understanding the problem, and work in front of someone while being judged confrontationally. It's not sane. (and glob help you if you don't wrap your entire brain around the problem in the first 5 seconds, then you're clearly subpar and are worth less than dirt even with your 15 years of experience shipping products to millions of users.)
Google is happy to have 1,000 nerd projects and only one thing that makes money. As long as they continue to interview smart people on technicalities instead of well-roundness and ability to actually contribute or move the company forward, they have no need to fix their insultingly irrelevant interview process.
Why not just take a chunk of code the candidate has written that was interesting and talk about it?
The main problem with interviews [imo] is it grabs stuff the interviewer is familiar with and throws it at the person in the more stressful position of being the interviewee. Its also [often] a very non-standard way of communicating. I don't know about you but all of my communication is via text or graphic, not whiteboard.
And, honestly, do you care more about the fact they can invert a binary tree from memory? Or that they can learn to in the next hour and implement it?
The point about interviewer vs. interviewee familiarity with the problem is important. It's hard to avoid comparing the interviewee's off-the-cuff solution to one that was researched at leisure and refined over multiple previous interviews. Really damn hard. Psychologists and teachers are trained in avoiding that kind of bias, and still struggle. Very few engineers are equipped to see through the "fog" and interpret a candidate's performance in a useful way. I've only encountered one - a recently former CS prof - and I've been in the business a long time.
I think the best way is to give them access to a code base, give them 48 hours and have them submit a pull request for a feature. That way you can see that they can learn the codebase and implement the feature in their own time. During the interview, you can discuss their code and the reasoning behind the implementation details.
I once got asked to spin off an ec2 instance from an image provided by the company I applied to, and then answer certain questions about the dataset on that thing. I had to write actual code to obtain the answers. All that had to happen within 24 hours. Then there was an on-site interview where they asked questions about stuff like JVM garbage collector fine-tuning and such. That was the best interview I had, ever.
Still I think that would become obvious pretty quickly. I've actually had people start writing out a textbook algorithm that just solved the wrong problem. And then completely stumbled when trying to explain how the program would arrive at the intended result.
I usually ask questions until the person is out of his comfort zone, and if he is completely clueless on how to proceed at that point it's a red flag.
Demonstrating one small part of a real working software engineer's job, by solving an artificial problem in an artificial environment under artificial time constraints, while one or more people with no pedagogical training look on and (sometimes) interfere. The amount of information gained is less than the amount lost from having poisoned the entire rest of the interview environment.
Number of times I have had to invert a binary tree in my 25+ year career: 0. Number of times I have been asked to invert a binary tree in an interview: 0. What I would do if I had to invert a binary tree: look it up.
It may surprise you, but in the real world very few people get paid to find new primitive methods of manipulating fundamental data structures. Google may do a lot of that, but they are an unusual company. The problems a working developer is asked to solve are far more likely to involve higher levels of abstraction. The one I worked on most recently was coarse geocoding (find city/state) of free text. But a couple of decades ago I did sort of invent a use of BSP trees for searching RGB palette space (http://www.drdobbs.com/cpp/vga-palette-mapping-using-bsp-tre...). Well, probably not, but I still think I should get a cookie.
>Number of times I have had to invert a binary tree in my 25+ year career: 0. Number of times I have been asked to invert a binary tree in an interview: 0
Well, if you wanted to pull that card, it would have been nice to mention the sorts of problems you've worked on in those 25 years.
>What I would do if I had to invert a binary tree: look it up.
Unless you're already good at algorithms, it would net you a mediocre solution.
For e.g. You would only know that there are multiple ways to implement an algorithm when you've actually done the work dozens of times and noticed that some implementations won't be appropriate for you needs. Like with many things, its about having experience, knowing whats a good fit, and knowing why something similar isn't a good fit.
Certainly - every single time you choose an algorithm, and then decide on a particular way to implement it - you could in theory, implement each algorithm in 10 different ways and then choose the best one after benchmarking. But that would be a huge drag on productivity. And if you had to choose 5 algorithms for 5 different tasks then it quickly becomes a quadratic complexity time sink. It would be far better to just know via experience. Its kind of like in chess - the better players just KNOW that certain paths lead to less favourable winning odds. Well, novices do take those paths and eventually find out the hard way !
>> Well, if you wanted to pull that card, it would have been nice to mention the sorts of problems you've worked on in those 25 years.
Quite a range of stuff, unsurprisingly. I started on an HP3000 mainframe in 1976, if you want to go back to the very beginning, writing BASIC programs on a teletype or one of the two early CRTs, and storing my programs on paper tape.
Since then I've worked in DOS, 16-bit Windows, 32-bit Windows, 64-bit Windows, Ubuntu, iOS, and Android, using Pascal, 8086 assembler, C, C++, C#, javascript, Python, and Java. I've worked on applications in multimedia, telephony, banking, insurance, pharmaceuticals, cosmetics, health care, and I'm sure a few other things I've forgotten.
All of which will mean jack squat if, tomorrow morning, the most important thing I have to do is invert a binary tree. But I'm fairly certain I would be able to figure out what I needed to do, and I am fairly certain I could manage to implement it well. It's what we're supposed to be able to do, and if you think that having studied up on it so that you could pass a Google interview means that, a few years down the road when you actually need it you'll just whip it off the top of your head, then I think life may hold some surprises for you.
Thanks for replying. I simply wanted to know what kinds of problems you worked on. A binary tree can be inverted presumably on any OS and using most languages. If you're going to claim that a basic binary tree operation has not be necessary for you in your 25+ year career, you should have mentioned your problem (!industry) domains. It was nothing personal.
>and if you think that having studied up on it so that you could pass a Google interview means that, a few years down the road when you actually need it you'll just whip it off the top of your head, then I think life may hold some surprises for you.
That is a misrepresentation of what I said. I said that _RETAINING_ basic and higher order CS fundamental knowledge is much more useful, in a way that simply looking up an algorithm on wikipedia would not be.
>> That is a misrepresentation of what I said. I said that _RETAINING_ basic and higher order CS fundamental knowledge is much more useful, in a way that simply looking up an algorithm on wikipedia would not be.
Sorry it was not my intention to misrepresent you. I was assuming that we all agree that simply looking something up without having any fundamental basis for understanding would not be of much use, and that we further assume the person doing the searching is in need of a refresh, and not basic education in the craft. In that context it is the idea that you might be called on to go up to a whiteboard and trot out something you haven't done in three years, and then be judged competent or not based on how successful the trotting is, that gets people worked up.
I totally agree with you. Even if a binary tree is considered a fundamental data structure I think CS is a large ass domain and you can go for years programming, designing data structures and working with algorithms without the need to reimplement basic operations on binary trees.
Another instance where silicon valley favors the young... he probably would've nailed this question if he'd been only a couple of years removed from school.
He also would've nailed it had he been given 10 minutes to do some research to refresh.
Silicon valley (and the numerous companies that mock the interview style) are testing for the wrong thing when they hire, then complaining about not being able to find good engineers.
You're given weeks to refresh your knowledge ahead of a Google interview. They tell you the basic CS fundamentals that are going to be covered. Binary trees are MOST ASSUREDLY on that list. They just don't have time during 45 minute interviews to let the candidate go on the Internet to look things up they should already have come prepared for.
Anecdotally, at a previous company, we tried an "open book" (i.e. you can use the web) interview policy for a few interviews. It was a train wreck.
I just don't get it why companies assume you can spend so much time to refresh and memorize topics just to prepare for an interview. Especially when this knowledge will only be used in the interview process and then forgotten about.
Not sure why you were downvoted, to be honest, because I think this is a reasonable take. I don't prep for interviews (or, these days, for sales discussions), except to read about the company and to consider how my skill set can help solve their problems.
Google's insistence on this preparation thing is doubly weird to me when I consider that literally nobody "prepares" for work every day in this manner; knowing what a developer can do after studying his brains out regarding things that will be ejected from the brain again immediately after the interview doesn't really send a meaningful signal.
>> You're given weeks to refresh your knowledge ahead of a Google interview
Begs the question, doesn't it?
>> Binary trees are MOST ASSUREDLY on that list
Sure they are... and my intro to CS textbook was ~500 pages. I'll bet I could find something to trip up just about anybody, especially when coupled with the pressure of doing it in front of the class ^H^H^H people that have $$$$/power over you.
At prior employers we've had very good luck with open-Google interview policies. I mean, we'd watch what you're Googling, and if you're copy-and-pasting we're going to drill you to make sure you actually understand it, but I expect to have a search engine when programming and I think you should too.
I prefer not to ask code questions at all, though.
It may be that we didn't give it a fair shake, as everyone who we tried it with probably would've failed anyway, but it turned what would normally be merely failing interviews into completely excruciating situations. When Googling is an option, a lot of people were spending all of their time in the interview attempting to cobble together vaguely relevant Stack Overflow responses. It wasn't just "I'm going to look up the exact name of this library function because I don't remember it", it was wholesale Googling of data structures and algorithms.
If all someone needs is help remembering the name of a specific function, then I'd rather provide them with that if I remember it or just spot them it as something trivially Googleable. I wouldn't hold that against someone unless it seems like they're completely unfamiliar with the language they've chosen to use. There's no need to give an otherwise good candidate the chance to self-destruct over trivial details like that by obsessively Googling everything to make sure it's exactly correct when that's not what I care about, and in so doing they aren't paying as much attention to overall algorithmic details, which I do care about. If they need more help than trivial Googling, then they can't solve the problem anyway, and me watching them thresh around on the Internet isn't helpful to anyone.
A good analogy is laptops in classrooms. A lot of students will attempt to use laptops to take notes, even though they inevitably serve as a distraction, and study after study has confirmed that students actually learn better when they take handwritten notes. Somewhat paradoxically, studies of classes where professors have banned laptops and forced handwritten notes show higher student happiness; they're aware on some level that they're more engaged with the class and are learning better, and appreciate that the choice to distract themselves is removed, because if the choice were possible, they might not be able to exist. Well, providing someone access to Google in a coding interview is very similar to letting students use laptops in classes, in a lot of ways.
A project manager at Google was upset that my 'bot literally ran circles around theirs (it was 2010, Android based robots were just starting) and told me that I was just a hobbyist and my project did not exist.
So I gave him one of the spare logic boards (open design anyway). And wrapped my hand around his. And squeezed. And asked him, if it doesn't exist, why is it making you bleed?
He watched impotently as the people who had been invited for the presentation played with my robots and ignored his as two of his guys tried to get it to work.
I finished the two projects I was doing with Google and did not call them again.
(Before you downvote: Yes, there is some video, and I consider the small amount of pain I inflicted that day a kindness compared to the much greater amount of pain that an engineer is in danger of enduring if he says things like "This thing that is in front of me, it does not exist", especially if he works with big machines).
I get that a lot. Very occasionally, it's after "thanks for saving my life / my job". There's no shame in being an asshole but there is shame and dishonor in being insincere.
No, a psychopath would've done something out of revenge.
I was doing something to remind an engineer that the physical world is of paramount importance.
A bit of bleeding and embarassment in 2010 as a reminder that things that are there, are indeed there, is better than being pulped by heavy machinery that "shouldn't be there" but is anyway in 2018.
I'm sorry but Twitter is one of the few (if only) megaphones most people have to call out corporations for their actions. This is exactly the sort of thing that Twitter (IMHO) shines at, a platform to call down the goliaths. If I didn't already want nothing to do with working at google I would have been pushed further away by this tweet which is exactly what should happen. If google has the right to say no to someone that wrote an extremely valuable tool that 90% of their employees use then he has the right to inform the general public. Popular opinion seems to be one of the few things that can cause a corporation to change their minds or at the very least acknowledge the situation and he is expertly wielding it to his purpose. I applaud and support him in making this choice.
Playing devil's advocate here (because I, personally, wouldn't have done something like that), but I can see the insult. Here's someone with a very public proven track record (that they're fully aware of, if 90% of their engineers use his software), and they're asking questions like this. They're the wrong questions to ask in this case, surely? It's fair enough to ask them, but I can understand this person's frustration if that was genuinely the reason he hasn't hired (which we have to take his word for, since we didn't witness it ourselves).
From his perspective, I bet he's frustrated for the reasons he point out. If he never talks about it out loud, how would any one know? Keeping people silent about this issue is in the company's interest because they'll not have to really change, and they'll hire someone eventually.
I mean, I know I sometimes choke on problems I'm trying to solve in the privacy, quiet, and comfort of my own home on projects I really care about.
I don't know what went on in the interview but given what he said, I think I would be upset, too.
Why? He's under no obligation to Google. It's not like they hired him and he has to sign an NDA or be polite. He had a frustrating interview process and vented about it online, it's a pretty human thing to do. Why would that make him a bad employee?
Well, not that I agree with the sentiment, but corporations likely have incentive to hire individuals that are easier to push over and keep quiet as long as they have the skill set needed.
If you were a corporation trying to keep PR positive, would you hire someone that has shown themselves as an outspoken public mouthpiece and tip-toe around them to keep them happy, or just avoid the problem?
If the tweet is accurate in its implication that he didn't get hired because he couldn't invert a binary tree (honestly, I don't even know what that means) on a whiteboard, then I'd say the tweet is totally justified.
On the other hand, if it was a complex and reasoned decision that he's summarizing badly, then yeah, good call on their part.
And I'm pretty sure this is EXACTLY the type of guy you want to hire, I would strongly suspect writing mild mannered complaints about not getting hired on twitter has very little correlation with job success.
Well, he is the author of numerous widely used open source projects. It is a bit redundant (and useless) to give him a "homework assignment when he has such an impressive portfolio that speaks for itself and of which you can assess the quality.
To be fair, though I know the situation in question: Company A wants to acquire Company B, because their product integrates well and also adds value to a particular niche of their customers. Person is one of the lead developers of Company B's product, and together they have also open sourced one of the single biggest components of their product.
i.e. "You want to buy us because of this, and a decent part of this is on GitHub, running every day, but you'd rather not look at -that- code but -this- arbitrary homework assignment".
He is doing both Google and people who were rejected by Google a favor by pointing out that Google's interviews are not indicative of real world performance. He might have hurt his own chances a bit in the process, but overall this will undoubtedly have a positive impact.
For him individually, I might agree, but the bigger issue is how one of the largest tech companies is hiring this way. I don't think it's worth focusing on this guy's understandable outburst to the exclusion of that.
As large organizations grow, their workforce trends towards mediocrity. Google
* takes special care to counter-act this effect.
* researches their hiring/interviewing practices just as much as their machine learning.
* publishes their methodologies: https://www.workrules.net/
I feel the original tweet conveyed a bad attitude, was emotional, reactionary, and ultimately a bad career move on the part of OP.
In my younger days I suspect I would have done something similar. I'd like to think I would see the experience as a learning opportunity and be able to react with humility and maturity, but who knows? Hopefully I can think of OP and not tap the tweet button.
Really? Pretty much everyone recognizes that google style interviews weed out perfectly good people. What Max went through is just an example of one such obvious case. It's very much a case of the google hiring algo failing. Lot's of people would have no doubt that Max can cut it when it comes to iOS dev.
That's all it is. Now, if you are going to read "tweet conveyed a bad attitude, was emotional, reactionary," into a perfectly human tweet, holy crap you are judgmental.
>> ultimately a bad career move on the part of OP.
Not at all. I've done a lot of interviews and basically none of them required us trawling twitter. I think it would have to be something pretty heinous for me to not hire someone based on their social media crap. Definitely not something as mundane as this. This sort of hilarious cowardice about expressing feelings just makes me angry. At what point do we stop acting like these trivial, humanizing glimpses into a person are somethign that is a bad career move?
I had a big comment here, but I erased it. I think one story proves my point. When talking to google engineers one thing I noticed was that they considered youtube to basically be a joke. The reason why is that youtube has a messy python codebase. I asked them what they worked on while they were at google. They had rewritten an internal web portal for a support tool. From everything I can tell it was literally a mysql crud app.
If this is how success and failure are determined at google, it's no surprise how many of their products that people actually use come from acquisitions.
I don't know if I would rant on Twitter, but I would be as frustrated as well. Those 'tests' are very academic and I am not an academic person. I've not even officially finished high school. This test will not properly judge how well I can code or design software. I've been doing just that for 20 years now, and launched a few start ups, but I would fail that interview at the door.
I interviewed once for Google (at their request) and failed. For some reason they interviewed me for a networking position instead of a code one, so questions about TCP internals were not really my forte. I was just launching my second start up at the time and would have declined the job had I gotten it. I admit it does sting a bit to be declined - not just for Google, but for any position - even those you would decline yourself.
As others said, much more than 10% of Google's engineers don't even own an Apple machine. So, even if Google's Mac management team somehow uses Homebrew to manage the employees' machines (which may or may not be true: I have no idea, even though I used a Macbook in Google until recently for years), the percentage of Google engineers using that software is nowhere near 90%. The percentage of Google engineers knowingly using that software is certainly closer to 10% than 90%.
I think there is some validity to the general point, but I'm not commenting on that.
Just quibbling: I don't think anywhere near 90% of engineers use homebrew? Google development is done on Linux, and Homebrew is a Mac thing AFAIK. I have never used Homebrew.
Android development can be done on Macs but I doubt they use Homebrew. Certainly not for anything important.
It's funny because the guy actually thinks having built something popular equals being a good software engineer. Wordpress, PHPMyAdmin and so forth are all really popular but the code is shit and though it's used by millions of people, a _real_ software engineer will shudder looking at its source code.
Now, I have no idea what the code quality of Homebrew is, but just because he built something popular doesn't mean he should get a green light in every company. If Google is looking specifically for top-notch software engineers, they are probably filtering them very well with their practices.
Maybe they are only good on paper at that moment and don't have something like "Homebrew" on their Github, their knowledge is sufficient to perform work at Google. So why pick someone who has fame to his name, probably wants to get paid accordingly and thinks he is a hotshot because of his Twitter and Github follower count over someone who proved himself in an interview?
The first is not necessarily better than the latter.
I strongly believe that the best kind of technical interview is to talk with the person about things they have done in the past, go into details and see if they are telling you bullshit. If the things are interesting, at least somehow relevant to what the company is doing and the person knows what they are talking about, it's a good hire.
One problem is that only experienced developers can do these kind of interviews, because you need wide experience, be able to talk about various technical topics and tell whether the other person is telling you stories from their own experience or some quickly learned facts.
It's funny, but the best experience I had interviewing at Google and Amazon was talking with the managers.
Was it really because he couldn't do the problem, or was it that he didn't handle himself well in the interview? At two different interviews I was given logic puzzles just so they could watch how I went about trying to solve them.
That experience doesn't mean that's just what they're saying. It could mean in those cases the person a) didn't solve them problem, and b) didn't demonstrate an approach to solving they problem they were looking for.
This just gets back to the question of whether you are hiring based on interview skills or coding skills. If someone has shipped code that shows high coding skills, why would you reject them based on interview skills, if it's a coding job?
You could still interview an experienced person. On the technical side, you could see if they're familiar with your specific stack, and know how to handle version control etc. Aside from all that, you can make sure they share the team's goals, are willing to use version control the way the rest of the team does, stuff that's not about skill.
That is the killer right there. The 23 year old recruiter responds to your rejection with: "The interviewer thinks you'll be better if you get another 6 months experience. Keep practicing! We'll loop back around to you next year!"
Thanks, I've been doing this for 15 years. I'm sure another 6 months is all I need to meet your delusional standards of ineffable ability.
Personally I don't really care about Google's hiring process. It's unlikely I'd ever want to work there anyway.
What does bother me is that other companies, who are not even in the same league as Google, start to copy their hiring process.
I remember interviewing with a digital ad agency a few years back and I swear, these guys thought they were Google. The number of academic trivia questions that came up, it was ridiculous.
In a way, I think Google has done a lot of harm to the industry in general by making others believe that everyone should have a hiring process like Google's.
I've been there as well. I interviewed at a design agency a while back. I nailed all of their puzzles pretty easily only to get there and realize I was going to be doing Wordpress hacking and other CMS work from 2010. I left after a few months because of how trivial I realized the work would be in the long term.
On the exit meeting it was relayed to me that they were having trouble finding people because no one could pass their tests and I was beside myself because I couldn't understand how they would expect someone with that technical ability to want to bang out Wordpress sites all day while there are 100s of people who would love to do that job and be very successful without ever even knowing what basic recursion let along the stuff they had in their test. Bizarre.
I don't know who said that, and while I strongly agree with it in principle ... in practice it just tends to not work out like that. Employers want a cog in a machine. They don't want you to 'stretch, grow or learn' on their dime - no employer is willing to take that risk. Further, career progression in the tech industry is commensurate with the degree to which you've been pigeonholed in a particular skill or task. So a high salary is usually indicative, not of the diversity of your skillset but how well you perform within certain narrow parameters.
i had a similar experience with ita software, later acquired by google. i'd submitted a solution to one of their puzzles (which was actually very close to their business) which exploited a symmetry in the data and my solution produced results significantly better than anything else that they'd seen (per the engineer that managed that puzzle)
interview was brutal - lots of whiteboarding very artificial problems totally unrelated to the business and i just couldn't get excited about it and didn't end up getting an offer
You just have to know the basic data structures and some algorithms for them. Liked list, binary tree, Hash map etc...
In the worst case, set up the data structure, then derive the algorithm. Do this in Java FWIW and make your life easy.
This is like knowing math and/or stats and applying it to a word problem.
Actually my pet theory is this: Interviews for experienced folks are more meant to keep them in their current companies rather than to filter the incoming new ones. This acts as a gentleman's agreement between the giant companies to keep their own talent pool semi-intact :) /s.
I mean imagine the amount of extra work someone has to put in to start interviewing. Take a break from your current work, Prepare your CS basics again, Prepare from interview question dumps online, read up / analyze everything the new company is doing and form reasonable opinions, practice white board coding / coding without an IDE, allocate time for any homework projects given, Psych yourself up if you are introverted etc. The alternative is just to stay in the current role and hope stuff gets better. 90% of the folks I know choose the alternative over the dehumanizing process of interviews. So many folks I know are good engineers get chewed up in interviews (both in my current company and elsewhere) because the process is pretty cooked. We are trying to see how this can be improved, but yeah - I just keep going back to my pet theory :)
I do agree with one of the commenters here though: At one point your resume should speak for itself. These are the kind of problems I would like LinkedIn to be solving instead of finding ways to spam me with recruiter deluge.
It's interesting the amount of hate and either rumors of bad experience or bad experiences directly. I interviewed for an SRE position last September and they were clearly trying very hard to make it a good experience no matter the outcome. I flubbed a couple of questions and they didn't make an offer, but the impression that they cared about my experience as an interviewee lasted. I wonder why my experience was so dramatically different from many here.
I've been invited to interview at Google three times. And they've declined to hire me three times. The last time I interviewed there the quality of the people that interviewed me was much lower than the earlier two times. I was still rejected, but I felt much better about working somewhere else.
I'm sure Google is still a great place to work, but its reminding me more and more of 1999 Microsoft. In fact the similarity is spooky.
Firstly, you're not entitled to any job you want just because you wrote Homebrew. If you accepted an interview with Google, then you accepted the fact that Google will judge you based on your problem solving skills, just like every other person was asked.
Secondly, I don't think this is a hard interview question; it's certainly fair.
Did you expect to be asked knowledge-based questions that Google knows you're already good at? Questions specifically geared towards you?
Or questions where Google can watch you solve a problem and be comfortable with the fact that you are able to solve coding problems?
Did you think Google would hire you to write Homebrew? Or solve problems on teams Google has?
That may be somewhat true, depending on how crucial and dependent Google is on Homebrew. But 90% of Googlers don't rely on Brew for work.
It is just a figure he made up to make a point about how popular his software is. Using his software outside the context of our jobs is no means to justify a hire. He should go through the same interview process as 90% (much higher than that actually) of Googlers.
My impression was that he was talking about usage for business, not personal purposes. As for the popularity, even if the 90% figure is made up, I was only trying to explain/justify his point of view.
It seems that almost everyone here knows better than Google how to hire employees for Google. Given that you can see why it is trivial to see through hours-long hiring committee decisions in just two seconds.
Edit: as pointed out by others, the hiring decision probably does not take a few hours, but under an hour. Still, the point is valid.
If their goal is hiring from industry, as opposed to from schools, I think I could credibly defend an argument that they don't.
(I think what they do now makes sense if they want the top of their funnel to be comprised mostly of recent CS grads, and I think that industry people with really high profiles, or that internal people really want, get to skip a lot of this evaluation.)
I have a lot of respect for Google (how could you not). That doesn't mean they can't be a bit clownshoes with recruiting.
"If their goal is hiring from industry, as opposed to from schools, I think I could credibly defend an argument that they don't.
"
I would agree ... if they were actually having trouble hiring from industry.
IE despite everything everyone ever writes here, google simply isn't having trouble attracting enough of the right people that they want to hire (despite arguments to the contrary that they aren't hiring the right people or whatever).
So i can't see an argument that they do it wrong as long as it's actually working.
I expect, if it's truly broken, at some point they will, but
I can't give you anything more than anecdotal responses to this, but Google is (a) missing (ie, failing to hire) specific excellent people in the industry, (b) cultivating a reputation that prevents people from considering them. There are bright spots, but my perception is that they work by circumventing/overriding the standard Google "backrub tickets for calibrated interviewer favors" process.
But yeah, I may not have been clear enough before: I do believe that what Google is doing is working for Google.
About the worst thing you can say for Google's recruiting as a business process is that it's somewhat inhumane, and pollutes the industry with faulty evidence (there appears to be a small cottage industry of consulting "porting" Google's recruiting process to other companies). But I mean, look at the rest of Startuplandia and Google's sins pale in comparison.
Well, this thread escalated quickly! Am I wrong in my understanding that when a company rejects they don't specify why and hence "rejected due to failure to invert binary tree" may be a guess here?
As Google bragged in 2013 that it managed a Mac fleet of over 40k and with a workforce of 55,419 in Q1 2015 (not just engineers, 2013 numbers were about 10k engineers), that's 72%+ of Google's workforce using Macs.
Homebrew is at least one of the best package managers for Mac. I would be very surprised if it was not at least near the 90% mark...
I mean - what's the point of spending 3-4 years in an Academic environment that proceeds to then test and grade students on exactly how good they are, at the time - then only to perform the whole process over again some number of years down the road, with fuzzier results?
Seems dumb to me.
I've worked with people who could likely do very well on algorithmic tasks - (of which most software projects require precisely zero) - but actually deliver something of use... not so much.
When I used to interview people (I wasn't a manager, but I was senior enough to be entrusted to the role), I'd just ask about projects they had placed on their resume (to get a feel for their contributions) and then the rest of the interview would be focused on what the job was and how that matched with their career goals, why they thought they'd want the job, that sort of thing. The latter part was a bit harder because people are naturally defensive during an interview, so they can't be like "well I want an entry level web developer job so I can parlay this into something better in two years" (which is, IMO, totally an acceptable answer), but you can generally politely get the idea. Maybe I'm weird, but I just sort of think interviews should be more about determining fit than giving someone a lie detector test.
I get the puzzlers or whatever for a phone screens (quickly weed out people that are obviously unqualified), or if you're hiring someone junior who doesn't have work experience, but if you're at the point where you're bringing someone in you probably think they're minimally qualified, so it should really be about determining if goals are aligned IMO.
I like the spirit of this, but he may well have not been hired for a variety of other reasons besides this -- including how he attempted to solve the problem or how he handled not being able to solve it on a whiteboard in an interview.
I hate whiteboard programming questions and I don't give them when I interview someone - I give them a laptop with 10 different languages on it, and some data to munge -- I think it's a pretty decent thing for both parties.
I, for one, agree with this kind of hiring process.
From my own experience - people that do well in such interviews are good generalists. On their own they will start discussing performance improvements and ways to parallelize the solution, it's a pleasure to have such an interview.
It's about enjoying problem solving and willing to keep your brain fit. It has nothing to do with memorizing solutions to some existing set of problems.
The people vilifying Max or saying "duh you didn't get hired, Google requires awesome people" seem to have a totally warped sense of exactly what Google engineers do day to day.
Blows my mind that there are so many people defending this (well-known and pretty much taken as a trade-off) lapse in the Google hiring algo and instead making it seem like Max's fault.
It is not. IMO it's a pretty ballsy and idiotic thing to do. That's partly why I am not too active on Twitter. I have a lot of controversial, unpopular opinions that I wouldn't want a potential employer to get a whiff of.
Google really does only hire individuals who are strong in theory.
Maybe we are jealous. We wish we had the brains of the people getting into Google. I can say personally that I envy these people.
The fact that they don't mind being treated as numbers maybe says something about these people too. They are cold. Their ego must be pretty big too if they make it into Google.
I remember a story from college graduation. A friend contacted an alumni from the university. The alumni worked in the semiconductor business. One of the questions he asked the alumni was: "How does he select a great employee?"
Alumni's response that it is exceptionally tough. He shared an anecdote about one of his best hires.
The alumni wanted to test the interviewee's knowledge in different areas. He asked a question on diodes - it might not have been diodes, but for the sake of the story let's stick to diodes. The interviewee replied: "Hold on, I am not one to know about it."
I have not worked at Google and I do not think I'd pass its interview process. It is unlikely that I be diligent enough to make Homebrew. Nevertheless, I am inclined to the idea, that being knowledgeable in all tested areas would not reveal the personal fit necessary to make a great team.
It seems that google wants to have code monkeys instead of creative software engineers. This is a common problem of big companies. IMO it is one of the reasons they get stuck with innovations.
Of course also the interviewers don't want to hire people which are smarter than they are because of their own career.
As someone who struggles to learn by rote as opposed to learning by practical means and has been both hired and declined by the Google recruitment process. I can't help but agree with his sentiment.
The recruitment process (at least for experienced engineers) should be little more then "can I work with this person". The 6 month probationary period that follows the hiring process should be used for "can this person do the job well". But that's just my experience, and it seems to have worked well.
Regarding the same academic questions everybody gets asked in every development interview, I feel Einstein said it best with "[I do not] carry such information in my mind since it is readily available in books. ...The value of a college education is not the learning of many facts but the training of the mind to think."
Google is about monoculture (certain type personality) - and from their business perspective it seems that approach works. Why to change it?
If they try to invent something new or in different market they might need different type of people but as of now ads business is cash cow and they would be crazy to try change it.
Without commenting on whether this is a good interviewing strategy, surely the point is to sacrifice some potential good hires in favor of definite good hires. In other words, you might be able to write pretty good code even if you can't solve problems on a whiteboard, but, given a choice, why wouldn't we just choose the people who can do that, too?
I think the stated philosophy of interviews like this one is that a false positive is worse than a false negative. Every single one of the responses to that tweet either misses that point or sounds like little more than defensiveness in the wake of a bruised ego.
You might disagree with that interviewing strategy, but you're not addressing it directly.
More or less agree with this. I'm also a Google reject (didn't make it past the phone interview). I didn't take the rejection personally. I don't see how the interview process could be drastically improved. They get a lot of applications and they need some way to filter - there is a standard and it has to be met. With the sheer amount of applications that Google gets it's a virtual guarantee that there will be a subset of people taking the piss regardless of what the interview process is like.
I don't doubt that the engineers who manage to jump through all those hoops are sensational. Personally in the end it just dawned on me that I didn't want to work at Google that badly.
The whole Twitter exchange is a pitiful sour grapes circle jerk, and I'm surprised that it's provoked such a massive response.
> why wouldn't we just choose the people who can do that, too?
I agree with your overall point (why not both?) but how does the interviewer know they can do both if they are only tested on their ability to solve problems on a whiteboard? I don't have a solution (that scales) to this problem but I tend to agree the types of whiteboard problems typically seen in interviews are a bad way to identify good developers.
This guy sounds like a tool. Sure, he's accomplished, but his website and linkedin are dripping with self aggrandization. "Splendid Chap" -- cringe. My hunch is that they didn't hire him because he didn't seem like a cultural fit.
How can we see if this technique works? There are two methods try to know something: deduction and induction.
First a little deduction. Let's try to be explicit about the theory behind this technique.
It's safe to assume that the job will NOT consist mainly of cranking out binary tree inversions on whiteboards while being watched over. So obviously we're hoping to make a correlation with something else. Assuming the candidate was not tipped off and learned this particular puzzle, perhaps we are correlating to an ability to rapidly create novel solutions of long-solved algorithms without reference tools.
But is that what the new hire will be doing? Probably not.
We could continue down this path, identifying ever more removed correlations until we get to something that the job actually demands. This probably involves solving hard problems like naming things. [1] But by now our theory stands on pretty thin ice indeed.
In any case, all of this deduction is theory making. It's not knowledge until we attempt to falsify [2] it via induction. The human mind constantly induces hoping to verify our deductions. We reason, observe, conclude and repeat. We're good enough at it to survive, but that's about it. Lucky for us, science came along. Today's technical hiring is at best alchemy.
A interesting company called TripleByte [3] is trying to apply induction (first for YC companies). They specially shun on white board coding and puzzle solving tests in general. I will be interested to see how they fare and whether their learnings are adopted more broadly.
It is pretty well known there are a lot of false negatives in the hiring process since it is so much worse to make a bad hire than it is to not make a good hire. Sounds bad and it is, but no one has a better solution than try again in a year.
IMHO, I don't think it requires any practice to be able to invert the binary tree. It is so trivial that it only requires a very basic level of programming skills. I agree whiteboard is generally broken, but for this particular case, I don't think Google is doing wrong. We can think another way, if some company hires people based on his reputation instead of the ability of doing actual work, I don't think it will survive. In this particular case, you just didn't show your ability of doing actual work, that's it. I am glad to see that Google prefers ability of doing actual work to reputation.
In my very limited experience? None. They just trust the company they acquired to have filtered you properly. They do ask for references (like academic records and so on) but that was about it.
Without discussing too many details, I believe the issue with Google's recruiting process is it was designed when the company was smaller and it follows the philosophy that anyone that goes through the hiring process should be ready to be thrown into any of the many Google projects and be able to function immediately. That's not strictly true anymore.
You have some divisions that are extremely hardcore or require very good knowledge of a particular field (think Google Cloud Platform vs. Android Kernel vs. Chrome vs. Search, all completely disparate projects), but there's also work for people that don't need to hold a PhD from MIT (think front-end development.)
Hmm, it depends a lot on the size/type of company and reason for acquisition. If it's closer to a acqui-hire where the employees of the "acquired" company cease development on whatever they were doing and eventually just get staffed on a Google project, then they will MOST LIKELY do technical due diligence on each team member. It's common for only part of the team to get an offer to join.
Good engineer you didn't hire is not much of a cost to the company (other than resources wasted on hiring process and perhaps some bad publicity).
On the other hand bad engineer will stay at the company, lower standards, damage morale and set bad precedent to other engineers.
Being engineer myself, I feel much more motivated working in an environment where you can just assume, even before meeting, that the other person is intelligent and motivated. You trust hiring process to filter everybody else so you don't have to subconsciously distrust every person you meet.
I dont want to be an ass but how do you not know how to invert a tree? Anyone who knows how to write a tree and traverse it should be able to do this. If you ran out of time coding it then that's different.
Not even. If you know what a tree is, and you've written a couple of recursive problems on trees in your life, then you know most of them are approximately 5-6 lines of code.
If you're spending 45 minutes writing 5 lines of code, it is not definitive, but certainly a red flag.
Nobody in this thread has even been able to define what inverting a tree means. (Reversing or mirroring? Sure.) My search for how to invert a tree led to a bunch of fairly hairy academic papers.
An Indian eCommerce company wanted to test my mathematics while interviewing me for a VP of Marketing role. I had to tell them I won't fit into their company culture, let's not waste time further.
We're trying a different approach at Sourcegraph. In addition to looking at a candidate's prior work in open source if available, we ask them to complete tasks that approximate the job as closely as possible (i.e., coding on a computer): https://sourcegraph.com/blog/programming-interview
Would love to hear people's thoughts and feedback!
Yours sounds like an approach that measures how well someone codes in a vacuum, instead of how they operate on a team. It very much skews the results...
Not to mention for a qualified professional candidate, it feels an awful lot like you're asking them to work for free.
In addition to asking them to write some code, we also have each member of the team interview them onsite to get a sense of how they'd interact as a member of the team. The challenge does take a few hours, which is longer than a typical phone screen or single onsite interview, but because it lets us focus on getting to know the person onsite rather than go through a gauntlet of whiteboarding interviews, we think it actually saves time for everyone and is a win-win. Obviously, every candidate is different; we think of this not as a rigid template, but a better default option than whiteboard interviews. Thanks for your thoughts!
> The code is reviewed by the interviewer
> but probably never run.
This is what surprised me during an interview with Facebook. I tend to write a little code, run it to test it out, write a little more, test again, and so forth, you know, rapid iteration or whatever the fancy industry term is these days. The interviewer gave me these silly little shell scripting problems (silly in that they ought to have been easy and clearly had nothing to do with real life work) but instructed me _not_ to run the code. How the heck am I supposed to know if I've solved the problem? How will I know if I even have the right approach? I don't consider myself an expert software engineer and so probably not a fit for the position, but that style of interview definitely hamstrung me.
For what it's worth, being able to run programs "in your head" is a very common skill required of Computer Science undergraduate programs, where you're required to write programs of medium length on paper during exams. You should be able to write out, follow along with, and reason about the correctness of a program that can fit on a single piece of paper without having to run it. Programming hyper-iteratively is not really a good thing, especially not in environments where rebuilds or test runs can potentially take hours.
Another reason not to have candidates run the code is that they tend to get really hung up with debugging trivial errors. I've conducted a fair number of interviews either way, and the white board interviews were usually more pleasant experiences for all parties than the interviews where the interviewee was expected to execute the correct solution in front of me. The latter almost requires giving them some kind of web or library API access, which then just makes the distractions worse. I don't want someone worrying about what the exact name of a sort function is; that's not what I'm trying to evaluate in an interview.
i've done three over the years, complete and utter waste of time.
i also had some wierd binary tree questions when i interviewed at yahoo, was a total turn off as i was interviewing for web/front end related job. which probably would not have used any of this type of logic, ever.
The problem with google interviews and tech interviews in general is that it is almost impossible to capture what makes a successful candidate in a couple of mini interviews. They don't even pretend that what you do in the interview is what you will be doing in an actual job there. Most of a developers time is spent in meetings, understanding their problem domain, writing documents, or reviewing other developers documents.
Succeeding at these whiteboard questions requires weeks of preparation. You need to practice, practice, practice. After enough practice, you are pretty likely to pass. So ultimately, it is more of a test of "how much do you want to work here." If you care enough, you can pass. That said, these types of interviews do favor fresh grads over experienced programmers (you have algorithms and data structures fresh in your mind), which means that they are flawed in IMHO.