Hacker News new | past | comments | ask | show | jobs | submit login
I don’t have time for coding challenges (css-irl.info)
121 points by PretzelFisch on Sept 11, 2020 | hide | past | favorite | 234 comments

> "I’m writing this as an interviewee, having never been on the other side of the interview table ..."

> ...

> "Most candidates will have side-projects or work they’ve done for previous employers."

As someone who has actually done a lot of interviewing, I can say with great confidence that this is simply wrong. At my current employer, we have a "homework assignment" as part of the interview process. We tell candidates that they can point us at an existing project of theirs instead of doing the homework. Few people take us up on this because few applicants have any substantial public side projects or FOSS work they can share.

And sharing "work they've done for previous employers" is a great way to court legal trouble. They shouldn't share it and as an interviewer I don't want to see it!

Them: "We've got a coding challenge for you..."

Me: "I've got this repo on Github with millions of downloads and 10k stars. It's at least 20k lines of my own code. Would that be good to demonstrate my coding ability?"

Them: "Haha, no, that'd be too much work for us. Please do the test instead."

Me: "Fine."

Then they rejected me because I used GET instead of POST. Clearly that shows I have no idea how to code.

That was a large (1000+ employee) YC company, by the way.

Truly absurd. Followed the link in your profile. You wrote VSCodeVim! 49,674++/14,291--. And because you used the "wrong" HTTP verb. Did they also say you should have used tabs instead of spaces?? Outrageous.

By the way, thanks for VCCodeVim!

Hah! Thanks :-)

> Then they rejected me because I used GET instead of POST.

OK, I'm honestly curious here. What was the request, and why did you make such a decision? Do you think it is the right approach in this case, or that it's an insignificant mistake that does not represent a person's understanding of HTTP protocol?

I remember hearing about at least one story where using GET instead of POST lead to disastrous consequences, as some robot crawled all over the "delete" links - but I have to admit, I can't imagine how it would happen without other significant design flaws happening as well (as "delete" links being public or robot being able to crawl under authenticated user's identity). This sounds like an error that can cause irrational irritation by the reviewer, but I can't imagine it being the sole reason to decide that the author can't code.

I was definitely in the wrong about GET/POST. The company I had previously worked for simply didn't really care too much about getting the verbs right, so I didn't really fret about them. And the company I'm currently at (multiple thousands of employees, strong engineering culture if I dare say so myself) just does everything as a POST. So it really is somewhat cultural whether you care about GET/POST or not.

But, IMO, even if you don't know, it's the kind of thing you learn once in code review and then don't worry about again.

For E2E encrypted APIs, one could well argue that worrying about HTTP verbs is just wasting time for no good reason. POST (consistently) lets you attach a body too.

Just using POST for everything is a pattern I've seen becoming more popular recently.

No we can get into an endless argument whether verbs are abused or not...

IMHO, GET (as of today) gas the issue that it does not support a body. If I‘m executing a search (an idempodent action), how can I nicely pass a JSON object configuring the search? I‘ll just use POST for that... So then we‘re at the point where the url does not work anymore because we have a POST /findEmployee; POST /findContract; ... instead of GET /employee and GET /contract.

Maybe I‘m wrong here, but it doesn‘t like nice in my eyes using POST in that way.

The JS API's I used to develop in used a mix of query strings (for exact matches) mixed with JSON (for more complicated queries) all in the url parameters

Google Places API puts the key in a query parameter. Still don't know how I feel about that.

Not the OP, but from my experience teaching web dev, one possible reason is nerves. I’ve seen some of the smartest students make that mistake on an assessment simply because they were under pressure. A simple conversation afterwards shows that they do understand the difference (as expected). In the moment though, sometimes you just have a brain fart.

This really can't be understated. The difference between your how your brain works when you're under the clock and know you're being observed vs. the "natural" flow of a calm working environment is tremendous. Anyone of us who's conducting the interview must do some kind of double-checking of obvious mistakes to fairly account for the limitations of the process.

Yep! I have been in similar situations where you make a mistake and you get out right rejected. Surely it's better to explain the mistake to me and then we can discuss it. This can also be a nice cultural check too, how well do I respond to my own mistakes? It's a shame most people don't follow your idea here.

> Haha, no, that'd be too much work for us

What's funny is that it's literally their job. They receive money for doing it.

Doing a take home is not paid. I have no idea why these recruiters assume it's ok to ask candidates for multi hours work without any kind of compensation.

That's absurd.

Also- thanks for VScode vim! I use it and love it. I've always wanted to reach out cause there's lots of features I wish would get implemented. Though even without those it's one of the better vim emulations I've tried, and I've tried many!

Seems like you are a brilliant programmer but lack domain knowledge for that specific job.

If you make that assumption based on a single thread comment, you lack domain knowledge on how candidates should be interviewed.

We use a coding challenge as well but I like your addition of "auditing" the coding part and using existing open source or side project work. Bonus points if it uses some of the target technologies too!

The author doesn't appear to understand the economics of the hiring company; it's about time & effort to filter and the cost of making a bad hire. I cannot give every application 1-2 hours of review if I get 100+ resumes. I can't give every interviewee a half-day assessment to get the very poor signal/noise ratio answers to questioning methods suggested. I can't hire easy and then fire easy. I can't invest weeks of work from myself and other mentors on non-ideal potentials. I know this means I pass over great candidates and that really sucks and is totally unfair. I'm sure that's happended to me as well. The problem is nothing presented here is a better alternative.

> I cannot give every application 1-2 hours of review if I get 100+ resumes.

This is perhaps the biggest misconception of hiring, that out of some misguided notion of "fairness" or "objectivity" you have to treat every candidate exactly the same. In theory it seems like a good principle, but the road to hell is paved in good intentions. In practice, this results in arbitrarily designed standard tests which are only superficially fair, and systematically weed out people who aren't good at test taking or auditions.

Go through all the resumes quickly, pick out the top candidates, and start from the top! Focus most of your time on researching your few top candidates, and if one of them seems good, then hire them, and end your hiring process. You may only have to interview 1, 2, maybe 3 people at most. This saves everyone time: your own time, and the time of the 100+ people who would otherwise be interviewed and rejected, because there's only 1 job opening.

Obviously if you interview your top candidates, and you find them lacking, then keep working your way down the stack. But going through the entire stack of candidates is a huge waste of time, a pointless exercise, totally counterproductive.

Don't do "screening" interviews. Only do "hiring" interviews.

Generally you don’t get all the resumes at once, this is literally the secretary problem. Unfortunately you need some objective standard of measurement across candidates that is static over at least months, and ideally takes a minimum of you and your candidates’ time. Though imperfect, real-time technical questions are a pretty good solution to this problem, especially since a lot of the best engineers prefer them to take-homes.

Also, anecdotally after interviewing hundreds of candidates, I’ve found resumes to have a relatively high false positive rate. Surprising numbers of engineers look great on paper but can’t write working code for semi-complex problems when tested, even accounting for interview anxiety.

Are you sure? I mean the 10 years of success spent over various companies projects on resumes have a high false rate because after they took your coding test they didn't do as well as you thought they should? Did you hire any of them and did they out perform your initial assessment.

If you didn't hire them and compare them to new hires who have bad resumes but did well on your test how would you know?

You can't objectivity judge a candidate by your test results only. Otherwise your test is just randomly filtering people. Until you do that you don't know how effective your test is. You could be flipping a coin and getting the same results.

I’ve hired people who have not passed the coding screen (but did pass subsequent interviews) and they’ve been some of my best hires. But the type of failure really matters. Did they write generally correct code but ran out of time because the algorithm they came up with under time pressure was too complicated? Or did they fail to debug a list indexing off-by-one error, even after printing the list repeatedly? It’s sometimes a judgement call but I believe there’s a lot of signal in watching people write code in real-time

> Surprising numbers of engineers look great on paper but can’t write working code when tested, even accounting for interview anxiety.

How have you accounted for interview anxiety?

It’s definitely not an exact science.

1. I try to make the parameters of the interview as friendly as possible — language of their choice, don’t have to worry about code style or computational complexity, or talk while they’re coding, just write working code that solves the problem. They can use google freely. I’m there as a resource to talk through their solution, answer questions, and help them get unstuck if needed.

2. The algorithm to solve the question doesn’t require a trick and there are many viable solutions so most candidates can verbally explain to me a working algorithm quickly, +- a couple edge cases.

3. Candidates often get visibly flustered or hung up on irrelevant details, or unsure what exactly I want for them, and I focus on noticing this. I try to guide them down a correct path if they’re close and tell them not to over-architect if they start doing that.

Despite all this, too manly candidates do not have a dependable metaprocess for writing working code - a common scenario is they write code, it gives the wrong result, they drop some print statements to debug, but can’t effectively trace the program to figure out why they’re getting the debug output and so they guess and try to fix the wrong thing, causing further problems. I know that some of this is nerves but it can’t account for all of it.

> I know that some of this is nerves but it can’t account for all of it.

Why not?

If you've never experienced severe situational anxiety, you may not know how debilitating it can be.

And then, maddeningly, perhaps just minutes after the interview is over, you return to your "normal" self, and start to think about what you would have done in that situation if you weren't nervous.

It's basically stage fright, which can be terrifying if you're not used to the stage.

I guess I'm generally going off of an interviewee's body language, which isn't entirely fair since I'm not a mind reader. And I'm sure I'm missing some cases of stage fright, which is why again I mentioned all the techniques above that I try to mitigate this, but I agree is a really unfortunate side-effect of the interview process that I don't have a fool-proof solution for.

Still, I do have some EQ. A lot of candidates exhibit visible anxiety and then get into a rhythm and calm down when they start coding. Or they visibly calm down once I start asking them questions about how they're thinking and coax them towards viable implementation. On the flip side, a lot of candidates actually exhibit overconfidence running in the wrong direction, which I try to lightly mitigate but can't prevent in all cases without giving an unfair advantage.

How do you "quickly go through" a stack of 100 resumes and choose a handful of top candidates in a way that is not subject to personal (conscious or unconscious) bias?

I remember an interview training exercise, which also covered resume screening. Each person got the same stack of resumes, and the instructor said, "You have 60 seconds to pick out the top 3 you'd like to bring in for an interview. Find them." At the end of the exercise we found that we all basically picked out people who were a lot like us. White people picked out "whiter-sounding" names, women picked more women, etc. Also, Ken Thompson's resume was one of them (with a different name) and nobody picked him.

I think the point was that human eyeballs and brains are terrible at objectively or fairly pre-filtering.

> How do you "quickly go through" a stack of 100 resumes

Well, by quickly I didn't mean only 60 seconds. :-)

> in a way that is not subject to personal (conscious or unconscious) bias?

In my opinion, a common conceit is that it's possible to design a process that is not subject to bias. It's not possible. There will be bias, no matter what process you choose. It's thought that "quiz show" style interviews are "objective", but if you look at the results, they're clearly not helping. They're not increasing diversity. The poor results speak for themselves. So then some people look at the results of this so-called "fair" process and conclude that minorities are just worse at coding, because that's what the tests say. To me, that approach and attitude is even worse than just acknowledging personal bias.

The only way to try to counteract unconscious bias is to be conscious of it. Admit it, and keep it in mind while you're going through your hiring process. Keep it in mind while you're filtering the résumés. Keep it in mind while you're interviewing. But don't try to pretend like you can eliminate it entirely from the process, because then you're just deluding yourself and everyone else.

Agree with this for most companies. When at an airport we were government and had to be 'fair' it was about ten solid days of work to hire for one position.

> Don't do "screening" interviews. Only do "hiring" interviews.

Good way to put it. Exactly.

This only makes sense if they want top talent. Usually they want cheap/easy to push around talent and for that you need a screener that filters out the best and whoever remains will have the strength to deal with bully managers or lack of requirements

The author doesn't appear to understand the economics of the hiring company; it's about time & effort to filter and the cost of making a bad hire. I cannot give every application 1-2 hours of review if I get 100+ resumes.

This is an entirely self-created problem, not for you personally, I mean for the industry. We are against "credentialling" so signals like personal recommendations, alma mater, or work history at prestigious firms are overlooked.

My employer tried Hackerrank for a while, we abandoned it when we realised that the candidates we really wanted (e.g. based on personal recommendations from existing employees) were dropping out of the pipeline as soon as the Hackerrank link was sent to them. And the ones we were getting were good at Hackerrank and not much else.

> I cannot give every application 1-2 hours of review if I get 100+ resumes.

Your process has already failed if you're considering 100+ people!

First, you should never be getting 100+ resumes for a role. If you are, you need to write better job descriptions and/or get a better recruiter who knows how to match to the job description.

With that in place, you (I) are getting about 10-20 resumes at most. Then what I do is read them and rank them. Takes less than two hours.

I'll only interview the top five from this stack. I'm happy to give each of these five people however many hours of attention it takes, they and I deserve it.

This way the people who didn't get to the interview round have not wasted any of their time which is the most respectful way to handle it.

> you need to write better job descriptions

Tech job postings are the absolute worst. Anyone who's having trouble with hiring in any way needs to start there. Everyone in tech knows that job postings are a joke, which is why you're told to apply to jobs even if you're not qualified, and then companies complain about the applicant pool. It's their own fault!

The "excuse" is almost always "HR". But this just tells the applicant your company is a bureaucratic nightmare that few sane applicants would want to join.

There's a lot of worry in tech hiring about the "faker", the engineer who lies or exaggerates on their résumé.

The irony is that the tech hiring process, especially the job postings, encourage the fakers to apply. You're practically beginning fakers to apply.

The most honest, scrupulous people tend to only apply for jobs where they're 100% qualified according to the job posting. So the system, which hires people who don't actually meet the listed qualifications, is systematically selecting against the honest applicants. Honest people can even be driven out of tech in virtue of being honest.

> Your process has already failed if you're considering 100+ people!

Remote roles get a _lot_ more applications than on site. In my experience getting 100 applications is quite common for such a role.

The commenter doesn't appear to understand the economics of the job-seeking candidate; it's about time & effort to keep your current job and the cost of getting hired at a bad company. I cannot give every company I apply at hours of coding if I apply to dozens of jobs. I can't give every company a half-day assessment. I can't get hired easily and then get fired easily. I can't invest weeks of work from myself on non-ideal potentials. I know this means I pass on great jobs and that really sucks and is totally unfair. I'm sure that's happened to me as well. The problem is that nothing presented here is a better alternative.

This is a snarky parody, but put together you are both just saying that you can only afford to spend any time on job matches likely to be especially good, but that you’ll spend a lot of time on the good ones. Which just means you both are in an illiquid or conservative hiring market that is out of your control and their control to change.

>I can't hire easy and then fire easy.

Why not? Maybe you should address that problem. It would certainly make things a lot easier.

This is on-point! I've been involved in the recruitment process at my current employer for years and 95% of people that make it past the initial recruiter meet-and-greet have absolutely nothing to show in terms of open-source contributions or personal projects. Most of them just do the homework assignment so that you can at least talk about something during the interview.

You can always say it's because the company doesn't attract good enough candidates, but good luck with that in a competitive dev market in a large european capital.

I appreciate the authors honesty in his obliviousness of being on "the other side of the table" and I think it's clear that the author is living in a bubble that doesn't correspond to reality for people recruiting.

fyi the author is female - https://css-irl.info/about/.

I'd say "work they've done for previous employers" is something that nearly every candidate with industry experience has. The problem is that when I ask about it, candidates often are terrible at communicating what the work consisted of: often they focus on microscopic minutiae that appears disconnected from the big picture (e.g. gloating about gluing some trivial library as if that's a big accomplishment, when the position is for a senior engineer) or they demonstrate apathy/lack of proactiveness (e.g. some variation of "I just did what I was told") or they ramble incoherently, etc.

And this is supposed to be one of those bog-standard question that interviewees are always told by job interview guides to be prepared for!

I agree that people that have OSS work to show are much rarer to come by, and FWIW, they tend to do better in interviews IME. My theory is that it's not that the OSS resume bullet point itself sets them apart, but rather that they polished their skills more through the hard work they put into their OSS projects.

Interviewers react badly if you try to describe the high-level thought processes you go through, because that's not "technical" or "coding". So you end up reaching for dumb bullshit like gluing some library or whatever the fuck, because coding is actually boring grind work that you want to do as little of as possible, as stupidly as possible, and there actually isn't anything interesting to say to these people if you did the engineering right.

If that's not you, then you need to be painfully, painfully explicit about it, and some people still won't believe you because interviewers just lie all the time.

Just to clarify, I thought the author of the linked piece was suggesting that people could share actual code from previous employers somehow. Of course it's reasonable to ask people about their past work in an interview and to expect them to speak about it coherently.

I think OP did suggest sharing actual work, which would make sense if said work consists of e.g. semantic html that can be looked at in a publicly accessible site.

Obviously sharing backend trade secrets would be a big no-no. I've had people say they can't talk about their work due to NDAs, and that's completely fine in my books. We just move on to another topic.

FWIW, even when a candidate mentions having an OSS project, I refrain from looking at the actual code, and especially doing code review. I don't consider an interview setting the appropriate place to be grilling people on why they used a while loop instead of a for loop or whatever, since a) nearly everyone writes "bad" code occasionally and b) often things are done in certain ways for some good reason that isn't really relevant to my goal of evaluating the candidate. At the end of the day, the interviewer's goal IMHO should be to get as many signals they can from the conversation and make a determination about whether the candidate meets/exceeds a minimum threshold of competence

Came here to share the same exact points. He’s making a lot of assumptions never being on the other side and actually interviewing lots of candidates. Most of the time I get half assed projects with bugs that are obvious and they haven’t ever made contributions to open source software as an alternative to the take home project.

I think the at home project is the most lenient way to test someone because they aren’t pressured to solve a problem on the spot, instead they are in the comfort of their own home and can reference whatever documentation they would need, just like they would in a real position.

We've done something similar and found even candidates that do have public side projects may not feel those projects fully represent the code they'd write today and preferred the takehome.

Good point. I'm sure some of the candidates who do our assignment have stuff on GitHub or elsewhere, but thought they could do better work with the assignment.

Then, don't.

I want to move from my 2000 USD a month (this is a good salary already) in a 3rd world country to a FAANG that pays 20k USD per month.

I'm willing to go through any challenge because the reward is superb for me.

Not to mention, without objective interview questions, an introverted non-native English speaker (who doesn't speak latin-germanic-based language) like me would fail most of these subjective questions (e.g. bad culture fit, can't articulate as well as a native speaker).

I like coding challenges. Please do more.

That's why people do it, no doubt. Keep in mind, a lot of these questions are asked by vastly less desirable companies that ape the hiring practices of FAANG, even though they pay much less. And these companies complain bitterly about a shortage of good developers to hire.

So... I'm all onboard with your statement, but I think you only said half of it. Nobody owes me a job, and if I don't like it, I'm free to not get hired. Nobody owes a company an employee either, and if candidates don't don't like it, the company is free to not hire. Zero sympathy for companies that do this and complain about a shortage of candidates.

> a lot of these questions are asked by vastly less desirable companies that ape the hiring practices of FAANG, even though they pay much less

I was asked to do a coding exercise for a tiny ($2500 if memory is correct) freelance project.

Do only those companies complain about a shortage of candidates? Because in the tech centers I know, the market is just empty. Lots of large companies are competing for talent, and if you can't pay bank salaries, you're going to have a hard time.

You make it sound like there's plenty of talent around but companies are being snobbish. That's not my experience at all in Germany.

In San Francisco, there's plenty of tech talent available at 500k a year. There's remarkably little available at minimum wage, though.

Complaining to congress that there is a shortage of IT talent is so common in American politics that it has become its own unique form of performance art. For instance, during the magic section of the act, companies make a labor shortage magically appear without any mention of wages, while also magically making the laws of supply and demand disappear.

“Shortage” is an economic term of art, so I won’t use it here. But it’s likely true that there is lot of valuable, economic-growth-producing work that could be profitable if a software engineer costs $100k/year instead of $500k/year. Businesses that lie on the margin of profitability are likely to feel that it’s hard to find good talent. Of course, it’s bad for us existing software engineers if the supply increases and the price goes down. It’s likely good for society overall though, if it enables a bunch of other businesses to exist. This is true in the same way that more trucks, more warehouses, and more conveyor belts would increase economic output, even if they were already available for the right price.

You just said that there's lots of valuable work that could be done, but your example of work that could be done is work isn't valuable enough to do unless software engineers get paid (in the SF Bay Area) notably less than a dental hygienist.

I like having clean teeth. If someone can make more as a dental hygienist, why on earth would it be valuable to society if they instead worked as a software engineer for less money?

If people won't work for 100k/yr as a software engineer, that means people capable of doing this work have found higher value things to do elsewhere. The higher salary elsewhere reflects the higher value produced doing something else.

This is one of many reasons why I'm so bewildered when people talk about a "shortage" of software engineers without considering the options available in other fields (fields that, perhaps, pay better and don't require taking bizarro whiteboard exams every time you change jobs).

I'd say there is a shortage of software engineers as long as most stem grads can easily raise their salary significantly by doing a few months at a coding bootcamp. That is still true, so there is a shortage of software engineers. I multiplied my salary that way, although I learned it on my own instead of a bootcamp but it is the same idea.

I didn’t say anything about the SF Bay Area.

That makes no sense. A voluntary pay reduction is like a subsidy leading to market distortion (more supply than the market can bear) and windfall gains (employers might just keep the extra profits) - only that this subsidy is coming from your and my paycheck lol.

The idea is an employer-side argument to counter the labour-side demand for higher wages: Don't have the wages to high, otherwise we go bankrupt and can't employ you any longer.

But programmers are already well employed, and software companies aren't really suffering. Cheaper programmers won't make website less bloated, won't make us figure out what to do with GPT-3, won't find a useful application for blockchains, won't make Windows 10 less moronic, won't make ad-tech less harmful, won't make the tech recruiters and the non-technical folk around software projects more competent...it would simply kill one of the few fields of employments that still pay well!

Look at construction. Did cheap immigrant labour make rents feel low in many areas? No, market forces drive up prices and owners keep the profit.

I didn’t say anything about a voluntary reduction in pay. Any reduction in pay would be a result of market forces in response to increased supply.

> Look at construction. Did cheap immigrant labour make rents feel low in many areas? No, market forces drive up prices and owners keep the profit.

In areas that aren’t constrained by density limits, homes are actually very cheap. In terms of square footage of home per population, the US is doing great. Problems are with inequality, in terms of geography and wealth.

Market forces should have already brought in more people into software if they really are that high, bringing down the salaries. What other measure would you suggest? Importing supply from abroad? Whatever it is, it would lead to a reduction in pay, which you suprisingly support - that amounts to walking into your boss' office and asking for a pay reduction.

> In areas that aren’t constrained by density limits, homes are actually very cheap.

That's the point. The effect of cheap construction labor only affects the rent prices in some areas, in other areas owners pocket the extra profit - while making construction jobs largely untenable except for immigrants coming from poverty.

I suspect something similar could play out in the software field.

> Whatever it is, it would lead to a reduction in pay, which you suprisingly support - that amounts to walking into your boss' office and asking for a pay reduction.

I just think it’s useful to try to understand the effects of policies not just on me, but on society at large. For example, I might observe that a tax increase hurts me personally but helps others. Are you saying that instead I ought to come up with reasons why the policies that are bad for me are also bad for everyone else, to better lobby for myself, or perhaps so I feel less cognitive dissonance?

> That's the point. The effect of cheap construction labor only affects the rent prices in some areas, in other areas owners pocket the extra profit - while making construction jobs largely untenable except for immigrants coming from poverty. > I suspect something similar could play out in the software field.

So in a region where home supply is constrained by density restrictions, there’s surplus extracted via high rents, and some of that surplus could go to construction workers (and they ought to fight to get a good wage so the developers don’t get all the surplus). Great for the construction workers, but bad for renters. In the analogy, software engineers are the construction workers and everyone else is the renters. Given the high market value of my wages, I will negotiate hard to get my share. At the same time, I can acknowledge that there would be benefits to society if there were more people qualified to write software, even though it would lower my personal wealth.

Currently there seems to be no mechanism increasing construction workers' wages or lowering rent. The extra profits are kept, and society isn't better for it (if it were, I'd be all for it like you). The reasons for this are fundamental, partly explained in an economics class, partly the result of imperfect markets and practicalities (a rich landlord might not immediately go out and build new real estate that would lower rents overall, he'd just sit on his boat).

My final point therefore is, let's keep IT salaries high for now.

In areas absent density restrictions, the US is indeed great at building tons of housing, developers don’t make huge profit margins, and rents are low. Is your argument that society would be better off if there were less construction workers, so housing costs would go up, allowing each remaining worker to be paid more and less total homes to be built? My opinion is that the higher wages for construction workers are nice, but they don’t nearly make up for the undersupply housing (we see both of these phenomenon in the Bay Area - higher wages for construction workers, who still can’t afford a home themselves).

You argued that software isn’t good for the world, so society wouldn’t benefit from more of it. This is a perspective I don’t usually get (especially here on HN) and I’m interested in hearing more. My belief is that software is part of a centuries-long trend of technology/productivity advances that have brought billions of people out of poverty, the continuance of which is necessary (although not sufficient) for a post-scarcity utopian future. I don’t know if we will ever invent a replicator, but I think more software engineers increase the odds!

Currently in NY making 150k with 10yoe at a non-tech company. Would jump ship in a heartbeat to SF for 300k, or 250k, or even 200k.

This legendary meme that everyone in SF is making 300k and drives a Ferrari has taken on a life of its own, and now we even have people outside of the Bay Area believing and repeating it. This is an outlier compensation for senior people at outlier companies. Anyone would jump ship for it, but it's not like hundreds of companies are handing out these salaries at the airport when you land.

Is it? Not sure about 300k, but 200k seems within the realm of realism. I've interviewed at a few Bay Area companies (I'd say they are "mid tier", not FAANG or top tier) and expected comp ranged around 200-250k.

That said, I was actually surprised at the lack of expensive cars driven by SFBA techies. I interviewed at Netflix, which I know pays astoundingly well, and the parking lot had a distinct lack of luxury brand cars. The exception seemed to be Teslas - I did see a bunch of Model S's and X's (Model 3 had not come out yet at the time).

Compare that to many East Coast suburban office campuses of non-tech companies that pay pretty poorly - you still see tons of mid to high end BMWs, Benzes, etc.

> Lots of large companies are competing for talent, and if you can't pay bank salaries, you're going to have a hard time.

Sounds like Economics 101. The curves of supply and demand meet at some point. Some customers can't buy (enough of) the thing because the price at the intersection of the curves is too high for them. Yes, that's how free market works.

There are things I can't buy, because they are too expensive for me. Or I could buy them, but I would have to sacrifice something else I need more. This is considered normal. When some company can't afford (people like) me, it's the same thing.

They would like to have cheaper developers. Sure, who wouldn't? There are also things I would like to have cheaper. Same problem. Why am I not entitled to get things as cheaply as I want, but my potential employer would be?

In my experience many companies prefer to whine about the shortage of talent, even if they could easily hire another person by increasing the offered salary by 50%. But they won't. Then they make millions of profit, which shows that the money was there, it was just their decision not to spend it on salaries. Okay, fair, it's their money, their decision. But if it's your decision, own it, don't pretend there wasn't an alternative.

Make it known that you offer double of the market salary, and you will be surprised how many people are out there in the "empty market". At least I never heard about a company that offered lots of money and couldn't find a person for the job.

That's all well and good, but that's not the point I was trying to make. Companies aren't having a hard time finding talent because developers are so annoyed by coding interviews, they're having a hard time finding talent because they can't/don't want to pay high salaries.

So developers go to work on finance and ad-tech, and not on normal, boring, only-making-three-times-the-national-average-salary stuff. And those companies then complain that they can't find developers. Because whether they have a coding challenge or not, they can't find developers because money.

I disagree with this, somewhat. Adam Smith, in Wealth of Nations, reminds us that there are a lot of factors that go into determining the desirability of a certain job, and that wages reflect this. The butcher earns more than the baker, Smith reasons, in part because most people would prefer the work of the baker. You have to pay someone enough more to convince them to be a butcher.

There are a lot of jobs that pay less than software engineer, but would be more desirable provided the pay isn't too much less. For example, someone might prefer to be a dental hygienist, figuring that the work is flexible, steady, and there's less of a possibility of age discrimination later in life. This person would become a software engineer for twice the dental hygienist salary, but not for 10% more.

In a free market society, it's not up to us to make that decision. If you want to get software engineers for in San Francisco for 10% more than a dental hygienist's salary, and you can't get any, that's the market's answer.

Corporate capitalists have discovered that if you don't like the market's answer, have the government ask for you.

> The butcher earns more than the baker, Smith reasons, in part because most people would prefer the work of the baker. You have to pay someone enough more to convince them to be a butcher.

And the surveillance capitalism developer earns more than the entire village. Ergo the market commands us to all develop surveillance stuff so Saudi Arabia can spy on dissidents and then chop them up in their embassies.

But again, that's not the point I was making, it's weird that multiple people understand my comments as "we should remove the free market". Me saying that it's the market that is making it hard for normal companies to find talent is only the background for what I'm actually trying to get across: It's not the coding challenge that is making it hard to find developers.

A company that isn't in a hyper-profitable space like ad-tech or surveillance (watching employees during WFH is booming) or fintech (figuring out new ways to make people part with their money) won't be able to attract talent by removing a coding challenge. It's just not an important factor in the equation.

I will engage with your claim that coding tests aren't making it hard to find developers. I think it is a more important factor than you do.

On a personal level, coding tests are a deterrent for me in applying for new jobs. I have my old copy of data structures and algorithms, and I've gone through "Cracking the Coding Interview." I actually enjoyed it at first. They're interesting and challenging problems. If there were an exam - maybe something similar to the actuarial calculus or probability exams - I'd probably be interested in studying hard for it and passing it.

However, I'm no longer interested in re-loading this information into short term memory in order to take a whiteboard exam. This may be a facet of how my own brain works vs others. I just don't walk around prepared to figure out which elements of a set, when combined after permutation, are equal to an existing member of the set. Or how to find all square sub-matrices with a positive determinant in a larger matrix. I can do these things - I've enjoyed doing them - but for me, it takes a lot of training and prep to re-load it in a way that I'm ready for the 45 minute whiteboard exercise. If it were a proper, lasting exam, I'd be happy to do it, but it's just too random, uncertain, and capriciously scored for me to want to do it anymore. I will note that some people don't have thae same trouble I have with this, and - I really mean this - this may actually indicate that they are better candidates than I am. In short, I need to acknowledge that I may certainly not be the "talent" that anyone is missing out on.

Money makes a difference too, but the reason I brought up Wealth of Nations is to draw your attention to the section "Of Wages and Profit in the different Employments of Labour and Stock" http://geolib.com/smith.adam/won1-10.html

I would recommend the entire article, as Smith breaks down the various factors that can lead to higher or lower wages, but you don't really need to read the whole thing, here's smith's summary:

"THE five following are the principal circumstances which, so far as I have been able to observe, make up for a small pecuniary gain in some employments, and counterbalance a great one in others: first, the agreeableness or disagreeableness of the employments themselves; secondly, the easiness and cheapness, or the difficulty and expense of learning them; thirdly, the constancy or inconstancy of employment in them; fourthly, the small or great trust which must be reposed in those who exercise them; and, fifthly, the probability or improbability of success in them."

Within this matrix, software engineering does have a lot to recommend it. For example, Law requires a very expensive and lengthy graduate degree, whereas software engineering doesn't, even though most successful SE's generally do have some kind of undergraduate, and often graduate, degree in rigorous fields that require a lot more math than law. How people will respond to this (along with other aspects of these two fields) depends a lot on their own personal wiring as well.

Within this matrix, yes, I do think that many people find in person, oral exams, at a whiteboard or at a computer under scrutiny and time pressure, on topics that they often don't know well in advance, evaluated by people whose credentials and skill they don't know in advance... yes, I think many people find that a deterrent, either for applying for new jobs, or entering the field in the first place. Many people consider these on-the-spot exams extremely stressful, especially when they are in person, at a whiteboard. And we go through days of it, and a new round of study, every time we apply for a jbo. Tech interview exams are not the only factor making it harder to attract candidates, but neither are any of the multiple factors identified and elaborated on by Smith. But I absolutely believe they are a factor.

And those small companies would be able to compensate the difference in pay with "partner" status, i.e. making an employee an unfireable stakeholder that can live off profits by the time he or she turns say 45, but owners of said small companies are too greedy for that. I'd argue that opportunistic at-employement with 500k/year at Big4 is worse than guaranteed 250k/year that nobody, not even the company's owners can take from you.

In my experience the talent shortage is a giant myth - there are two thing going on: companies that can't hire or won't pay, and companies that can, but wish it were cheaper.

Evey job I've applied or hired for either a) had lots of candidates with a very selective process, or b) telegraphed that you should avoid the company or didn't pay well enough (and often still got plenty of applicants who wanted their foot in the tech door despite not being capable).

My current company would claim there's a shortage, but they actually don't know how to hire and telegraph that you probably don't want to work there in a variety of ways.

What you describe isn't a shortage. It's an unwillingness to pay the going rate for the talent they want to hire.

All the more reason for candidates to not do the coding challenges. If you can't find candidates, don't add stupid ways to reject candidates.

As someone who had administered dozens of "objective" coding challenges to non-native english speakers... I can guarantee you your language skills will negatively impact your performance. So those tests aren't as objective as you think.

It comes up with questions that are as simple as "Create a function that deduplicates an array of integers"

> I can guarantee you your language skills will negatively impact your performance.

As it should.

If someone is a poor English speaker, at an English-speaking company, they absolutely should be penalized.

Not everyone is equal, was equal, nor will be. Some people are born blind too. Shouldn't we adjust our hiring practices for them too?

> Some people are born blind too. Shouldn't we adjust our hiring practices for them too?

Not only should we adjust our hiring practices for blind applicants, there's a legal obligation to make "reasonable adjustments" for these candidates under the Equality Act! *

* Laws may vary in less developed jurisdictions.

Well, it's more objective than asking a subjective question.

> 20k USD per month

As someone who's been in a FAANG for nearly 8 years and is making 2/5 of that...where the hell did you find that position!?

EDIT: oh wait. I was comparing my take-home salary. You're probably describing gross total comp? In which case - yeah, that makes sense. Good luck!

20k/month is only 240k a year. It's a take home TC for an average senior engineer in Bay Area.

Dude - you're not even making $180k/yr (gross) at FAANG and you've been there for 8 years?

What is your position...? I earn about $9k/month (net) at $170k/yr (gross) in CA as a software engineer. So, I'm already past the 2/5 mark and I'm just at a startup that doesn't even pay well! On top of that - you should be getting stock that translates directly into a good chunk of cash - not just a salary component. (Unless Netflix I guess)

Check out https://www.levels.fyi/

Check my edit! I was talking about my actual takehome salary (post-tax, not counting RSUs). Total gross comp is just north of $300k, which is pretty much in line with levels.fyi's average for that position (a little below - which tracks. I've only been promoted into that position just under a year ago, so I'm probably under-average across the whole population, despite being in the Bay Area so I'd hope to accelerate past that average soon).

I don't know of any software company that would hire an introverted person who can't articulate themselves for 20k a month. This is how micro-service architected fully optimized solutions that solve the wrong problem happen. Do you really think the high value of a developer just comes from being able to pass coding challenges when there's millions of them out there?

Eh, each of FAANG will do exactly that (as will other tech companies!). To be charitable to the parent commenter's point, they probably mean articulate in two senses:

1. Speaking perfect English, both fluently and without a heavy accent.

2. Talking about technical topics or other shibboleths with the same cultural affect an American would, as opposed to a foreigner. For example, many Indians will call a phone screen a "telephonic interview", and I've personally seen Americans react poorly to that.

I know lots of developers who are very well paid and very effective in their jobs, but they don't speak English as well as an American and would stumble at certain cultural nuances.

Thank you for being charitable. I speak English well enough to fare in FAANG as senior engineer.

But I always notice how a native English speaker almost always speak perfect English. They pick the right word to use and many other small things that make their English superb.

You seem to be confusing english speaking ability with the ability to communicate and articulate well. Communication is not about being able to speak english perfectly, it's firstly about listening and understanding the context of requirements. It's about being approachable enough so that the clients/stakeholders want to work with you, and clarify things with you, which in turn leads to a product which is close enough to something that they needed.

I don't think they're confusing anything; I thought their use of the word, "articulate" was pretty apt.

No worries :) You speak English well enough that I agree you shouldn't have any problems communicating in a technical discussion. Personally I find thick accents to be a bigger obstacle to communication than stumbles in grammar. But realistically most engineers will be professional and charitable enough to work around either issue as long as you're earnest.

That's only 240k probably not for a base salary but total comp sure.

It's a total comp. I'm not sure why we don't talk about TC especially when it's a public company.

Well, I have a bridge to sell you if you believe it's total comp.

I've seen messy devs who absolutely can't articulate themselves, to the extent that when they tried, everybody else would keep awkward silence, and yet they made 60k/month.

What do those devs do to make 60k/month?

720k TC is hard to attain but possible.

Any decent engineer joining Facebook in 2012 and sticking there until now would earn 60k a month.

Isn't that because of the stock which just happened to appreciate that much, not because Facebook wants to pay them 60k / month?

Yes, that's correct.

You're complaining that, to get a job, you have to demonstrate your ability to do the job, in a standardized way so the employer can compare you to other candidates with the merest hint of objectivity.

For a little (imperfect) contrast, here's a quote from an article about auditioning for the Baltimore Symphony:

> In the audition I took in October of 2014 for the Baltimore Symphony, I began preparing the required audition list in August.


> so the employer can compare

And why do we think this is objective? Ostensibly, the first one to clear some threshold of competence ought to be good enough to be hired (in fact there's even extensively studied algorithms[1] for this stuff). In the vast majority of cases, it's highly unlikely that you're going to find an underpriced genius via some misguided notion of test standardization.

[1] https://en.wikipedia.org/wiki/Secretary_problem

Big companies aren't looking for underpriced geniuses. Standardized hiring is like index funds, many argue they can get better results by hand picking stocks but index funds still tend to outperform managed investing unless you are really really good at it. In a big company the people doing the hiring will be mediocre at it, so you need a process they wont screw up.

I think it's objective because I can weed out candidates based solely on objective questions about their solution to the coding problem. Of course there's still lots of subjectivity in the process, but I did say “the merest hint of objectivity”.

Here's my real world case study.

I wrote a programming problem for RGM Advisors. The company was sold and so its web site no longer exists. Here's someone's solution to that problem:


My original problem statement is in the PDF in that repo.

Here are some of the objective ways we evaluated submissions:

    - Does the submission parse/compile?
    - Does the submission run and not crash?
    - Does the submission follow the specification for how it should be invoked?
    - Does the submission produce the correct output?
    - Did the candidate submit answers to the time/space complexity questions given in the problem statement?
    - Do those answers correctly characterize the submission?

> with the merest hint of objectivity

That's all it is. Superficial objectivity.

Auditions are not a good analogy. Programmers are not stage performers. We're usually sitting alone in front of a keyboard. That's where we do our best work. Not with others standing over our shoulders, holding our entire fate in their hands, ready to kick us out the door if they're displeased in any way in the moment.

Yeah this is tough I agree and I've been on both sides and felt this pain.

On the bright side a very excellent means of hiring and finding new work is through colleagues you've worked with in the past. So do good work and keep in touch with past people you've worked with.

> You're complaining that, to get a job, you have to demonstrate your ability to do the job

That's not generally what people complain in these interview threads. The complaint is that to get a software job one must demonstrate a performance skill 100% unrelated to the actual job.

It's as if a surgeon with 20 years experience in his specialty would be interviewed by making them balance chemical reaction equations. Sure, they studied that back in undergrad prior to med school but it's completely irrelevant to the job. But that's how we treat software interviewing.

Lots of software interview practices would never fly in the medical world:

Take home tests, work samples, and coding challenges are like asking a surgeon with 20 years experience to remove someone's appendix at home as proof that they are competent.

Whiteboard coding is like asking them to remove someone's appendix on the spot with a 30 minute timer running while a bored other surgeon pays attention to which tools he or she uses.

There are tradeoffs and "objectivity" may be overvalued.

I'm of the opinion that if there needs to be some objective technical challenge to determine a baseline of competency (and there are very many, very strong arguments for that) then there need to be standards.

I've encountered so many variables (no pun intended) in technical interviews that I've virtually stopped preparing. There's hardly any way to prepare when the field is so large and you never really know what you're going to be tested on.

I've had people challenge me on theoretical system-time timezone conversions (and incorrectly mark my correct answer wrong "because of physics"...), beside reimplementing native JavaScript functions, beside complex algorithmic problems, beside toy problems like building a binary calculator. [That's great if you learn I can rig a web page for adding 100 and 101, but how do you know if I know anything about databases for this back end position]

And those interviews were all for working with web tech, in a couple of the cases just front end positions.

There needs to be a common, limited standard and/or domain-specific standards.

It's really counter-intuitive to me, and I don't understand how it benefits the hiring process, if a company is looking for a web-back-end engineer for what is essentially a complex crud app and wants to propose the candidate solves a problem from a domain they've never worked in—and marking them on the correctness of the answer, rather than the work process.

I'm for objective, technical challenges. But I've seen enough that it seems apparent to me that it's broken.

(addendum: I've had several good interviews as well, and good interviewers. I had a good one recently as well that was fitting for the domain. They're not all bad, but most seem to be in my experience)

I interview a lot of people for frontend positions. I ask one two-sum question to see if they have some knowledge about data structures and simple algorithms. Then I ask them to build something trivial in React, which is what we use. If you can do those things, you're probably fine. The number of people who can't is truly shocking.

I would argue there are many well-qualified frontend engineers who won't correctly solve the two sum problem on their first time encountering it.

This one? https://leetcode.com/problems/two-sum/

>>> Given an array of integers nums and an integer target, return indices of the two numbers such that they add up to target. You may assume that each input would have exactly one solution, and you may not use the same element twice.

Seems trivial enough to me... except if the interviewer forgets the hard constraint that there is exactly one solution, then that's a shitstorm of edge cases.

I'd criticize the problem for asking for the positions of the numbers rather than the numbers. That's just super annoying for no reason.

Does the problem actually have that constraint? If so then yeah it's significantly easier. I thought the reason you'd have to use a clever algorithm is because a brute-force solution to find all index pairs would require n^2 operations, where n is the size of the input array, because you can't stop after you find a single pair.

Technically, if all values in the array are the same value X, and the target is 2X, you can't do better than O(n^2), because just writing writing all valid solutions is O(n^2).

If you can create a new array containing indices to the original array, sort the new array, and use two pointers starting at the beginning and the end of the sorted array to find a solution, it takes O(n×log(n)) time. It is not necessary to assume that the solution is unique, assuming it is okay to simply write one of the possible solutions.

If all values are between 1 and k, and you can make a k-sized array, you can find a solution in O(n+k).

No matter what you do, you can't get better than O(n).

The important thing is that there is more than 99% chance you will never have to solve this problem in your job; and even in the case you would, there is 95% chance that the O(n×log(n)) solution would be acceptable regardless of whether it is the best possible or not.

Yes it does, you can open the leetcode and see for yourself. That also explains the "easy" difficulty rating.

I guess you've been giving candidates a more hardcode version of the problem, not having that constraint?

edit: nevermind, it's the other poster who was giving that problem, not you. No idea how they formulate it in their interviews.

I don't ask candidates these kinds of questions in my interviews. For a coding interview, I'd normally ask them to implement something like backend rate limiting logic for an authentication service, and continually ratchet up the complexity on each successful implementation until the end of the interview.

I was asked the two sum problem in an interview once, but I forget if it had this constraint or not. This seems much more reasonable.

Forgive me for my ignorance, but are you looking for anything more complex than a leaky bucket algorithm + gossip protocol solution?

If we're being very charitable I guess the idea of using a HashMap for an optimal solution might slip your mind but even a freshman in CS can just loop twice over an array and solve it.

I might be misunderstanding the question; how many pairs can there be? You can't just loop over the array twice to find all qualifying pairs of indices, if there are multiple. You'd have to loop over the array n times, where n is the size of the array. Isn't that why the question is difficult? There doesn't seem to be much incentive to using a hash table, complexity wise, if you could just solve it in O(n) time and O(1) space by naively iterating over the array.

He meant "write two loops" when he said "loop twice". It seemed obvious to me that he meant that but I can see how if you're used to more unambiguous language it could be confusing.

Yeah, you just do it in O(n²) worst case because the problem is O(n²) to find all pairs that match because there can be O(n²) pairs.

However he meant one pair here too. Pretty easy to solve. Did it in under 30 s in Python on my broken cellphone. If there are multiple pairs, just replace return with yield.

Same here. Absolutely shocking how many people fail at basics, yet they pose as mids or seniors even. I'm not even talking about CS, I'm taking about simple React-based assignments designed to show you know (at least) the basics and takes as little as time as possible to be fair from our side not taking too much of interviewee's time. Take what you want from it, but I've seen it mainly in web dev related positions.

Yes, I see people fail at basics in interviews all the time. But I keep wondering how much of that is people being poorly qualified, and how much of it is interviews being poor indicators of true skill.

I've seen people barely sneak through technical interviews, whose subsequent performance justified the concerns I had at the time. I've seen others hired against my personal recommendation who turned out to be brilliant at their jobs.

Of course. That's the primary reason I'm personally against tasks on the interview itself or whiteboarding. That's usually a shitshow. We first do an interview where we talk a bit about our company and the interviewee, we get to know each other out a bit and try to find out if we're a good match for each other, based on the talk itself. If that's the case, we give a small (really small and basic, unless a senior position) assignment to take home and get back to us in five days. People have schedules, obligations, we understand. During the period we're always available for any questions, doubts, whatever you need. Assignment itself takes a couple of hours at most for someone that knows / has used the tech. At home, you can use google, stackoverflow, friends, books, I don't care. No one cares, as long as you show up with a consistently-looking code that works and that we can comment together on. You know, same as at work. Something to talk about for a bit, if at all. It's the basics. A damn good filter, it turns out.

> ...if there needs to be some objective technical challenge to determine a baseline of competency...then there need to be standards

After having been on both sides of the table, I’ve really come to appreciate the system of licensure used in the trades. I don’t have to quiz an electrician on the electrical code or ask them to wire up a dog house. I also know they have requirements to keep their license active. I’m not sure what that process could be like for software, but it would sure be nice to eliminate the enormous collective waste of time spent (poorly) establishing candidate competency (especially when they have to do it for every single company).

the FAANG interviews are pretty standardized now - but then that only selects candidates who are well prepped for the interview itself.

They're really not. It becomes a cat and mouse game where the interviewers know which pool of questions people are studying so they try to invent new ones to prevent people just memorizing the answers. The questions thus end up becoming more and more obscure.

yeah but they're variations on the same stuff. it just allows gaming of the whole process where if you solve enough of those types of questions, you have a better chance of getting in.

Some are variations but some are just more obscure. They definitely get into very specific greedy algoss or things like prim's/kriskals/tarjans. which can be pretty much hit or miss based on if you've seen it or not

> that I've virtually stopped preparing

Surely this is a good thing, no way for people to game the system and no time wasted.

Oh no. I've wasted plenty of time. And so have they by interviewing me and proposing ludicrous problems well outside of the problem domain of the role I was interviewing for.

I understand some problems are meant to judge your thought process, others are meant to measure your basic literacy. I'm not talking about those, however.

My experience in interviews has been that most of them are engineers pulled into the process who searched the web for "software developer interview questions technical" and cobbled together a few disparate options without a moments concern more.

Then again, those kinds of interviews have certainly cemented my desire not to work there.

Work sample tests have been proven repeatedly to be the best predictor of future job success.[1] The next best predictor is general mental ability.

It's not unreasonable to ask someone to demonstrate their capabilities. Just make it interesting, relevant to the job and most importantly: short. Whatever time work sample testing requires should replace interview time, not add on to it. 2 hours is a reasonable ask. Oh, and you can't ask people to invest time in a work sample test unless you've already invested a significant amount of YOUR time on in-person interviews. [2]

screening interview -> coding assignment -> maybe more interviews? = not cool

screening interview -> first round with hiring manager -> coding assignment -> review assignment with candidate -> maybe more interviews? = cool


[1] https://www.researchgate.net/publication/232564809_The_Valid...

[2] Me. I'm a ex-Netflix software engineer, now run a tech recruiting company (www.facet.net). I've seen hundreds of technical interview processes across many industries.

Tech Sector Job Interviews Assess Anxiety, Not Software Skills https://news.ycombinator.com/item?id=23848039

The sample size in that study was pretty small.

How good a predictor is a GitHub account with open source contributions?

I would argue that a GitHub account with open source contributions qualifies as a work sample test. It's just work that was done in the past.

The best test predictor from the research I read is prior work performance at prior companies.

The proposed solution to me, is much worse.

> Ask them to build something – anything!

> Instead of a coding challenge, ask them to spend a short time building something they enjoy, and talk about what they did and why.

To me this would be a MUCH more difficult task. A blank canvas is much harder to work with, and downright frightening. Constraints foster creativity.

Having everyone code the same thing helps the interviewer (who is also very busy and has many other things happening in their work life) gauge the different levels of the people applying easily by comparing them on the same task.

The difference is the interviewer is being paid for their time, the interviewee is not.

That's a problem with the company. We at @cdr pay interviewee's $200 for their take home challenges.

This is awesome. I have taken a number of code challenges and walking away with neither, compensation nor explanation is rough.

Then pay the candidate! It might seem ridiculous but consider how much time is spent by interviewers, hiring committees etc, giving some fraction of that out even to utterly worthless candidates might not be the worst idea in the world. A side effect of this is that suddenly the cost of a bad screening becomes more intuitive

That would create the wrong incentives.

"What do you do for work?" "Oh I just interview a lot, the pay isn't amazing but it is what it is"

heavily dependent on the job as well. not all software engineers are tinkerers solving basic problems in their lives.

I code state of the art motion planning, localization, and robotic algorithms. some of this requires more than a day of tooling around at home building a project

I've done 5 coding challenges, I deemed each one a success on my end, none resulted in a job. A few resulted in a third interview which revealed numerous additional red flags.

At this point I consider a coding challenge a deal breaker. I'm not doing one, and if you ask, there's a near certain chance I wouldn't want to work in an environment that thinks they are a good idea.

If you're interested in my abilities, I'd be happy to spend an hour walking through something I've already worked on in the past.

I agree that code challenges aren't ideal. But we've been struggling with parsing resumes and titles to identify the competency of developers. Our current coding challenge isn't super hard and based on the type of code you'd see in our product. It asks you to read a simple javascript function, explain it, and then implement new functionality that can built by refactoring the existing method. It's about 20 lines of code and 50 lines of documentation. You can use any IDE that you like.

We've had many candidates that look strong on paper (3+ years of experience, senior titles, etc) come in and struggle. Struggles ranged from basic code literacy (think for loop to traverse a nested json object) to basic programming like writing a function. We've especially noticed this more with coders coming from purely a javascript and front-end background (React/Angular, etc). They understand control structures like if/else but struggle with loops, setting variables, and basic higher level concepts like breaking a problem into smaller functions.

My philosophy is for junior candidates, any level of competency is ok. Strong mentoring and experience on job usually can allow anyone to thrive and so you can prioritize soft skills and culture. But for more advanced roles, I don't know, it's hard to parse out the quality candidates without having some sort of technical interview. Definitely would love to hear how other's learning and experiences.

I've heard a variant of this repeatedly. A company justifies technical exam style interviews on the grounds that many people who look good on paper "can't code". Meaning, can't program fizz buzz. Can't write a method or function. Can't write a loop, or an if/else conditional.

But I can do all these things. I've gotten bounced from my interviews because (this was the official reason, anyway) I didn't get far enough with the coding in finding all matching subtrees in a binary tree, in 45 minutes, at the whiteboard.

Companies should feel free to hire based on my ability to do this, they should not say they were just testing basic things like lists, loops, and conditionals, or even basic data structures and algorithms like binary trees or hash maps. These exams go way beyond that.

This may not be the case for all, and it sounds like it is not the case for your process. But in the broader industry practice, I smell a motte-and-bailey fallacy here.

In our case our its based on our actual interview experience with candidates. We've been actively evolving our process and calibrating expectation based on the candidates that have come in. We've actually simplified the interview question considerably (as I mentioned the solution doesn't even require you write new logic, you can refactor existing logic that's provided). But we still see candidates struggle with the basics.

To be fair this could also be a reflection on our ability to read resumes and experiences and our initial phone screen. But it is hard. My first job was at fortune 50 company, you could have the title of senior software engineer and not write a single line of code as part of your job.

That being said i don't won't diminish your experiences. This is just our experience from the other side of the table. Also we don't have an HR department (we're a small startup). The interviews are conducted by engineers in the company and we try our best to empathize with candidates and be conscious of creating an interview process that we'd feel comfortable being a part of as well.

I don't see how this addresses geebee's point -- what makes you so sure struggling on programming puzzles implies incompetence on the job?

we're not asking programming puzzles. i described our process above. struggling to define a function and basic code comprehension (e.g. tracing a for loop that traverses a json object in candidate's preferred language) is a strong signal that candidate is not a good fit a mid to senior level for us. If it was a junior level position, sure you can learn on the job.

in our case, the problem is a highly simplified version of functionality in our code base and is close to being representative of work they'd be doing if hired.

i actually don't disagree with the prior poster. Technical interviews are highly variable and it's frustrating to go through poor interview processes. I've got bad interview stories as well. Being on the other side of table now, we're always interested in learning about better approaches hiring. Would love hear what has worked well for others.

How do you know that?

If you're interviewing someone for an 'advanced role', you can talk to them and find out if they know their shit. If you can't then you shouldn't be interviewing them. Senior devs can see where other devs are at pretty easily.

For advanced roles I'd suggest separating out the evaluation of coding ability from the evaluation of a particular language. Let the candidate pick their language when you want to test how well they can write new code, and then do a separate evaluation of how deep they are on the language and technical stack you are using.

I agree with the sentiment. Honestly we don't check for syntax or even require the code to compile. We let the candidate write puesdo code which usally isn't super far python and javascript implementations of our challenge. We try to pay attention to the types of questions the candidates asks about context of the code and choices they make in designing a solution.

That being said i'd be curious how you evaluate general software engineering ability (e.g. writing clean abstractions, decomposing monolithic methods, and general design). I feel most system design evaluations are a bit too high level or overly broad (e.g. design a library application or make a social network) and coding challenge are too low level.

I like to think of the higher level as creative problem solving.

Are they (me in this case) going down rabbit holes? Are they painting a clear picture? Are they communicating that picture? Are they suggesting edge cases? Are they minimizing bloat? Is it too complicated? What are the most important aspects of the design in their eyes? Can they argue for them? Do they understand where a piece fits and if not, do they understand that something has to go there? Can you both step back and definitely outline a logical flow?

There's not always one solution, but the approach counts.

Lower level is logical problem solving.

Can they reason the problem space? Do they start naively? Can they iterate? Can they optimize? Can they explain each of these steps for future iteration? Do you understand what they're writing? Would you work on that code yourself?

There's not always one solution here either, but 90% is getting a naive solution that can be optimized by taking a second look. I feel that the naive should be found in 5-15 minutes and you can spend the rest of the time talking about how to optimize.

I know I'm rehashing what you probably already know, but at a certain point interviewers are just playing games.

If you've found someone you want to work with who can also reason about problems in a timeframe you see fit, well... that to me looks like gold.

I like these questions because that's how real systems are going to be built.

Take-homes are great, but there is just so much effort wasted in the process if the product can't be understood.

One of my previous managers would set up the white board interview for a candidate by giving them a simple problem application (not a course-grading system but similar) to draft out on the whiteboard an initial design. Then they would add in additional details/requirements in about three stages. So the candidate at some point would typically have to back track and revise some of their initial design to accommodate the new requirements.

It was a good measure of how well they communicated, how they dealt with undoing some of their prior work, and the process was fairly representative of the type of work involved in the position.

Another we'd love to hear opinions on. Based on a recent study where candidates are more likely to succeed if given private space and time to work on the problem, we'll disclose the coding challenge about 15 minutes prior to the interview allow the candidate to read it before joining the call. We struggle with the right amount of time provided for private review is. We want to be fair to all candidates and not implicitly favor candidates who have more free time to dedicate to the interview process than others. but at same time provide some sort of reflection space outside of a cold start to the challenge.

I like the coding challenge we designed for hiring where I work. A few things about it that work:

- It's designed to take about the same time as the round of interviews it replaces (half a day), and we clearly communicate how polished and complete we're expecting it to (not) be in that time.

- It's a lot like the actual work; if you enjoy it / are good at it, there's a good chance you'll enjoy / be good at the job. It helps make sure (on both sides) that the job is a good fit.

- But at the same time it's designed to be obviously not actually work we're going to use; we're not exploiting unpaid labor.

- There aren't gotchas to it. We review it with an eye toward "if my coworker showed me something they'd been working on for about this long to address this problem statement, how would I feel about their work," and do our best to express that.

- The rest of the application process takes the challenge into account in a positive way. I might have a conversation with you later to understand how you think about the problem you solved, but none of the interviews have to be gotchas or on-the-fly skills tests because I've already seen the kind of work you do.

- Everyone takes the same challenge, and it feels way more fair to discuss candidates based on reviewing how they do the kind of work that we do, compared to "X person has a cool open source repo; Y person gave a good interview answer to this technical question; Z person worked at an interesting place in the past." There's still some apples to oranges in comparing coding challenges, but much less than comparing one person's bootcamp group project open source repo with another person's reason that their best code is closed source but it sounds cool.

With a well-designed challenge the benefits on the hiring side are really obvious (we actually have a basis for comparing applicants that feels somewhat related to the work they're going to do), but also (I hope) we've managed to make a better process for applicants without expanding the time they're spending.

a) Do you compensate folks who take these challenges. Otherwise it is half a day of unpaid labor. Also most take home tests take way longer even though they are advertised as "only a couple of hours". Do you have empirical evidence that it can be completed in the given time slot? Have you given the test to your own colleagues who may not know the background / domain knowledge / environment setup required?

b) Do you make it clear what you consider as "good design"? OOP/Imperative or Functional coding style? Even in OOP, is it sufficient if the candidate just does SOLID or do they have to implement GoF patterns? This is very subjective and never made clear in these tests.

c) Is there a guarantee that an actual person will review the submitted code and provide feedback? AFAIK, most companies won't give candidates any feedback / validation of submitted work

Not OP, just a job seeker who like coding interview.

> Otherwise it is half a day of unpaid labor.

The same true for an interview

> Do you have empirical evidence that it can be completed in the given time slot?

As a candidate once you see the specification and you think it's not worth it you can easily can say no to it.

> Is there a guarantee that an actual person will review the submitted code and provide feedback? AFAIK, most companies won't give candidates any feedback / validation of submitted work

That is why I only willing to do these kind of challenges once I had the first round of technical interview and I know I would like to work there

> Otherwise it is half a day of unpaid labor.

>> The same true for an interview

Who does half day interviews? In my 3 years as a professional software dev, I've perhaps spent a cumulative 2 hours in interviews with a company. I know that some companies have extensive interviews. But they aren't as laborious as intensely grinding on a mentally challenging project in an industry such as software.

I like coding challenges, they way better (less stressful) than whiteboarding. Just don't use it as a prescreen. Interviews (should) work both ways, coding challenges only testing the candidate's abilities. I'm only willing to do them once I know I want to work at your company.

What I love (sarcasm) is companies that just need basic line of business apps, but want you to complete a coding challenge navigating graphs and trees in constant, nlogn time, etc.

I find it hilarious and often ask, "Can you explain to me how solving this challenge applies to what you do day to day?"

They say.... "That's a really good question..."

Then I watch them squirm as they have no idea how to explain a real world application beyond a generic, sometimes hard problems come up...

Sure... Then give me an example.

I don't mind taking your test, but if your team doesn't need those skills you're wasting your time and mine by cargo-culting this time-consuming style of recruitment.

how about "we found a strong correlation between who can complete coding challenge do well on the job. We understand that correlation doesny imply causation; but we and the majority of the industry havent found anything better yet. We also understand that we are potentially missing out on good candidates who dont generally fare well in coding challenge; but thats a trade off we are ok with"

If you are FAANG and you have sea of candidates applying for any job position, that's fine. But if you are a small company please don't cry afterwards that there's not enough qualified developers out there.

“Great, thank you for your time! I hope you find the right candidate for your team”

> how about "we found a strong correlation between who can complete coding challenge do well on the job.

But that would be a lie. So not a good way to start a (potential) working relationship.

How did you find that correlation?

Majority of companies including FAANG (not sure about Netflix) do these type of challenges

I require them because candidates lie, not always intentionally.

The unfortunate reality is that it can be really tough to fire low-performers. I NEED to know that a candidate can do the work. We go out of our way to make our ask reasonable, relevant to the role, not spec-work, not time consuming, and well defined. It's not uncommon for us to learn that the candidate can't actually do what their resume says they can or for the candidate to learn that they grossly overestimate their own skills.

I recently talked to a Python/JS expert that led a project at a major agency to scrape a network of websites, find and extract specific data, and deliver game changing competitive intelligence. The candidate ended up admitting that their role was limited to copying and pasting a pre-built code snippet into the dev tools console and that they couldn't actually read or write any code. The proof is in the coding challenge.

I find these people reveal themselves almost immediately when we start talking about their projects in any depth.

Funny how he starts saying he never had a whiteboard interview and he never gave an interview, yet he has such strong opinions on everything that's wrong with it.

edit: read to the end, a lot of the points are absurd, easy to realize after a couple interviews.

Hiring a bad developer can be a costly mistake. In addition to the costs for the developer, there are often other costs and productivity lost.

Basically a good coder is not the same as a good developer. Coding is relatively easy - if you can visualize and solve problems using code, then you're a good developer. It's the thought process that makes you a good developer.

Coding challenges are often a good way to test this thought process. I often require candidates walk me through their code and explain their reasoning behind everything to see if they'll be valuable on my team.

However, coding challenges are just one way to ascertain someone's thought process. Watching people work on code during a game jam or hackathon often gives me an idea of their thought process, as does the advice of other people I trust who vouch for them.

So, no - coding challenges aren't always necessary. But unless there is a better way for me to learn whether you're going to be a good developer, then I'll give you one.

The author's suggestion of a 'trial period' of employment might also be one of the impractical, costly, infeasible ideas I've ever seen proposed.

So the candidate doesn't want to spend 1 hour on a coding interview because 'they don't have time.' The solution?

Spend a month on-boarding the candidate to your company (taking time from multiple team members), give them an implicitly high stakes work project while they get used to the new working environment and production codebase.

Meanwhile there's tension/friction because the company wants to quickly validate the temporary hire. If it's successful, okay proceed as an employee. If it's unsuccessful, it's a huge waste of everyone's time, a costly experiment for the company/team, and likely a very unpleasant 6 week long working engagement. Then the candidate will have to explain why they only lasted at their previous employer for 6 weeks in their next set of interviews.

The author's suggestion of a 'trial period' of employment might also be one of the impractical, costly, infeasible ideas I've ever seen proposed.

Then the world of 1099 IT “technician staffing” and ‘contingency technical recruiting’ will absolutely horrify you because an entire cottage industry unto itself is built around this employment model.

It’s especially popular in call-center workforce management.

This person has admittedly never even done a whiteboard interview or given an interview. Why are we upvoting this drivel to the front page?

Boohoo, you have to spend a couple hours proving you know data structures to get a new job. Give me a break.

Let us assume a random candidate Mary who is batting at 20%, i.e., she passes to the next stage on every 5th attempt. Now if we follow the typical interview loop

a) Take home test b) Phone screen c) Onsite d) Offer

How many take home tests does Mary have to take to get 1 offer? The answer is 125 and if each is 2 hours then 250 hours. So a single test in isolation may not seem like a big deal but if every company had their own criteria and take home exams as a pre-qualifier, the candidate would face a uphill battle to find work.

I mean, my takeaway from that is that Mary is not a very good candidate, not that there is something broken with the interview process. She only needs so many interviews because she has a low success rate. Most people aren't doing nearly that many interviews in an actual job search.

That's why even though I like take home tests I only willing to do them after I had the first round of technical interview. I always pass these type of test, that's why I like it at the end of the pipeline.

> Have a probationary period

That seems very taxing on the company. I'd rather screen the candidate before committing to anything via interview/coding challenge, etc... The take-home coding challenge seems like a good idea -- less pressure than coding during the interview and allows the interviewer to assess the "real life" performance.

> Ask them to build something – anything!

That's great, but a big part of working at a company is building something that you don't like or don't understand well -- talking to stakeholders, customers, figuring out specs, etc -- something that a good take-home exercise can assess

The probationary period seems incredibly stressful for the employee. If I am currently employed I would never switch to a role that uses a probationary period as a significant evaluation step because of the risks of leaving my stable employment, and not passing your probationary period. As an employer I wouldn’t want to limit my selection to people that either are unemployed or have a higher personal risk tolerance to handle such situations.

Exactly. Not to mention younger engineers will probably work 24/7 for 12 weeks while constantly worrying they could be fired tomorrow because they didn't fix some bug fast enough.

Not to mention it eliminates anyone who already has a job from the candidate pool. No company I know would be OK with you also working full time at another company simultaneously.

> That seems very taxing on the company. I'd rather screen the candidate before committing to anything via interview/coding challenge

So having a probationary period is too expensive for the company, but a several hours coding challenge for the developer is a OK?

In my experience when a candidate was a bad fit, it was because of soft skills/personality traits not technical chops (which are much easier to learn than soft skills). Are you screening for soft skills in the coding challenge?

The developer is eventually joining the company, no?

I didn't like spending probably about 5 hours total on coding challenges last time I interviewed, and I've been internally opposing them, but as the candidate who interviewed, I'm much happier having dealt with it once and gotten over it than having probationary employees show up on my team for a week every so often and having to drop what I'm working on to work with them and evaluate them.

> In my experience when a candidate was a bad fit, it was because of soft skills/personality traits not technical chops (which are much easier to learn than soft skills).

This is a good point, and argues in favor of traditional interviewing (establishing a rapport with the candidate etc.) over any sort of "challenge" IMO. I don't think I'll learn more about soft skills in a week than I would in an hour - the complex stuff generally takes several weeks for the new hire to be properly onboarded and contributing, anyway.

This sentiment is too real. And too many folks in "charge" of hiring play this game.

When I'm hiring, I give applicants a real-code task, something they'd really do if they were hired - and I pay for their time too.

This method has allowed me to skip a lot of time-suck-bullshit while finding candidates, and their process with me is easier/faster too.

> When I'm hiring, I give applicants a real-code task, something they'd really do if they were hired - and I pay for their time too.

Most real code task require a lot of ramp up time to gain context.

Agreed. We use a coding challenge that mimics the type of work we do, and we pay people for their time. We talk through their solutions and get a good sense for their ability to focus, follow written user stories / specs, and their familiarity with common tools. We also get a feel for how well they can talk about their work and the product.

We see a strong correlation between performance on this task and performance and happiness when working with us.

The other important point is that this is very similar to the work we do and the way we generally do it. If someone doesn't like or want to get paid to do this task, they're likely not going to be happy doing it every day if they get the job.

I think one can be creative about it. Applying the principle "nobody wants to buy a 2 inch drill bit; what they want is a 2 inch round hole", the people in the company do not want to give you a coding test, they want to hire a competent candidate. So convince them you are competent for the job in other ways if you do not want to spend the time doing the coding challenge.

One way of going about about it is to find former colleagues that work for the company you are applying to, and ask them to provide references on your behalf.

I interviewed somewhere that offered 3 options for the technical part of the interview. The options were a take home assignment, whiteboarding, and walking the interviewer through a non-proprietary project. I thought that was a great way to make the interview process more fair.

Time is money.

If you make about $100 an hour, a six hour coding challenge costs you about $600.

If a potential employer asked you to spend $600 to fly out and see them, on your dime, would you do it?

I would do it IF I desperately needed a job. I did it back during the dot com crash; I needed a job badly and I was willing to roll the dice.

In 2020? Oh hell no. There's a million companies who are hiring, and an employer would have to be exceptionally special for me to waste $600 on the chance they'll hire me.

Companies fly people out all the time (well, pre-COVID). They pay for the travel expenses, but not your time. I personally find this much more of an imposition than spending a few hours writing some code in the comfort of my home.

The only time I've bothered with coding challenges is what I was out of work, and zero of them, even some I spent hours working on, have ever lead me to a job. My current employer was fine with an example from current open source project I contribute to.

I would never work somewhere that didn't make me write code as part of the interview process.

I don't want to work with people who are good at bullshitting their way through interviews, but can't actually write code.

This is definitely something to consider. I actually once turned down an offer because I thought their process was too perfunctory. (Along with some other signs.) I didn't like what it implied about how they saw me as an employee/resource nor what it implied about my potential teammates. (And since the process was so minimal I also didn't get much chance to vet the team/working environment.)

And I don't have time to hire you. You wouldn't hire a visual artist without seeing their portfolio. I don't hire a software developer until I they can quickly write a program that compiles read and follow a spec, name variables well, take feedback and ask relevant questions. To not test candidates for these skills would be recklessly negligent for my employer.

I freelance, and if asked to do a coding test I'll tell them to pay for my time, usually they drop the test and contract me right away.

I think that the coding challenges can be helpful if done correctly. If the challenge is clearly communicated on how complete they are expecting it to be and how it might be judged. But I did have an experience with a "homework assignment" that hit on most of the bad points outlined in the author's article.

This was not specifically a coding challenge as I was applying for a Systems Engineering role at a software company. The homework assignment was basically here is a fake product, please provide a requirements spec, user manual, and test procedure associated with the product.

My first problem was with the notion of doing the systems engineering backwards and creating requirements specs from a product.

The second was with the conflicting communication about how much time and effort should be put in. The actual homework assignment itself said "We anticipate that this exercise should take no more than a few hours". Whereas the email I got from the recruiter said "They are looking for extreme detail here, and the ‘it should only take a few hours’ instruction is kind of BS. It’s OK if you don’t get it back for a week or so. Winning submissions from my candidates have been 7 and 11 pages long, I believe. They really want someone to go overboard with the exercise".

So after putting a lot of work in (because I could, and did not have a family at the time), the only response I got was basically "Nope, you didn't get the job". And I know that employers have a lot of candidates and there is a disproportionate time spent by the employee versus the employer, but it would have been nice to at least receive 1 or 2 sentences of feedback that may have helped me in the future.

To a point made by someone further down in the comments, the exercise may be a good way to judge if you would be a good fit at the company. So based on my experience I am probably glad I didn't get that job.

My employer carved out a standalone PHP script from a dark corner of a legacy application years ago. We show it to candidates and talk through what they'd do differently. I like it because I can judge not just the candidate's technical competence, but their level as well. If you're a junior, I can see if you have an eye for detail and pick up on the database query with a syntax error or mistyped variable name. If you're more senior, you'll tell me about job queues and workers. I had one candidate start with "the problem is the management structure that allowed this to happen in the first place". I have plenty of leading questions for candidates who are hesitant to criticize our code, and will tear it to pieces in front of them if necessary.

And no, most importantly, it's not a take-home exam.

As someone who has been programming for 40 years and had a decently successful career but is struggling with these challenges now, I can empathize.

It's not so much what the challenge is -- it's how one is asked to accomplish it.

Sitting in front of my laptop for 6 hours while people ask me random programming tasks while they watch me on camera and see every keystroke on screen while I'm editing is not how I work. I work by choosing from active tasks that fit the "space" I'm in, by multi-tasking, etc. - if I get stuck on something in the "real world" I'll go work on something else that needs to get done, etc. You'd be amazed at how many interviewers simply refuse to ask you another question in lieu of or even in addition to the one the first wanted to ask, for example.

> Have a probationary period

As someone who's done a lot of interviews, my rule is that letting someone go during their probationary period is an absolute last resort. It'll never be used as an excuse to take risks during the interview stage - except in extreme situations, everything should be decided up front.

This is purely for the benefit of the candidate. Many people quit their old jobs only after they've received an offer for a new one. Some even have to uproot their lives (and that of their families) and relocate. Firing someone during their probationary period for a preventable reason seems negligent. I'd rather put them under a bit more stress up front during the interview.

Why not just ask the candidate if they're willing to do a probationary period rather than deciding for them?

I suspect the type of people I'd need to fire during their probationary period are the ones who would be desperate enough to say yes to anything during the pre-offer stage.

Also, firing people isn't something I really enjoy, so even just for selfish reasons I'd rather avoid it. If it's preventable, I'd rather just get it done property at the start, rather than take these risks later.

And if it got to it, HR would make my life difficult.

Without having a degree, I feel coding challenges are the only way I can truly compete.

Many of the author's suggested alternatives clearly fail on the stated reasons why coding challenges are bad. Asking pertinent questions is clearly arbitrary; probationary periods are non-inclusive to candidates who are relocating or need a visa; asking candidates to "build something - anything" would be substantially more time consuming for engineers in many specialties. (How am I supposed to build a small demo app to show off my systems programming ability?)

It's hard for me to escape the conclusion that the author's just trying to universalize their personal preferences.

I think the best hiring process would give the candidate options. Choose among pair-programming a live coding challenge or a new feature creation, whiteboard, take-home coding challenges, or going through some open-source project from the candidate.

I happen to prefer take-home challenges more than any of the alternatives. It’s much less stressful than any live assessment. And I have the time.

But I understand people that hate it. You evaluate the candidate how they think they are best evaluated. And I don’t think it is that much more trouble or time investment by the company.

I get the general feeling that people on both sides of an interview want it to be exactly what they want it to be. That never happens. There are problems with the way tech interviews are conducted now, and we should aim to fix it. But human biases will always apply just like every other thing in life.

My rule for interviews (whether as an interviewer or interviewee) is simple:

1. Explain/ask for the exact process and stages during the first call.

2. Raise any concerns as soon as possible.

3. Be ready for either outcome - if your concerns are reasonable and addressed, go ahead. If not, don’t.

I've been on both sides of the interview process and I don't agree. If you were a chef applying for job it wouldn't be unreasonable if an interviewer actually asked you to cook something (one famous chef would ask candidates to fry an egg).

I do, however, not expect any candidate to do anything we haven't had to write ourselves. I also let them look up Google/StackOverflow etc during the interview. They're also offered a beer but nobody ever takes me up on that.

I have my own misgivings about asking interviewees to write code for interviews. But some of what the author considers weaknesses of this are actually strengths:

> A disproportionate amount of time is often spent trying to deduce the requirements of the challenge.

Deducing requirements is a job skill at least as valuable as the actual coding! There's nothing worse than perfectly polished code that solves entirely the wrong problem.

And the author's proposed alternatives seem even worse:

> Ask pertinent questions.

OK, but that selects for people who can build a compelling narrative around their work. This is not the worst skill to have, but it's not entirely aligned with the actual work requirements, and subject to cultural and sometimes gender biases.

> Have a probationary period

I strongly dislike that. At-will employment is a reality in many places, and probationary periods are a reality in even more places, but they should be seen as a last resort, rather than a routine filter. I don't want co-workers to show up and be kicked out within a few weeks. I don't want to move 8 time zones, enroll my kids in new schools, sell my house, to be told after two weeks "Hmm, this does not seem to be working out".

> Walk through their existing projects

As has been mentioned numerous times on HN, that's at least as discriminatory as take-home exams, if not more so. You may spend your free time with your kids, or welding kinetic sculptures instead of programming. That's a perfectly valid choice. People who interview forensic pathologists don't expect them to cut up bodies at home in their spare time.

And generally, showing existing projects from previous employment should be, and is, grounds for immediate disqualification.

> Ask them to build something – anything!

That strikes me as considerably harder and considerably more subjective to evaluate than a standard take home exam.

> Pair program / Focus on their ability to learn

That's reminiscent of one technique I used to like: Show the candidate a piece of product-like code, and ask them questions about various pieces in it (programming language constructs, algorithms used).

I'm currently in the interview process with few companies, here's my experience with them:

Company 1:

1 hour HR interview

Coding task spend about 4 hours on it, at first i implemented basic but correct implementation, no OOD & design patters or anything fancy. They said i should redo it but with SOLID OOD and design patterns + unit tests, so i threw in the most convoluted code i could think of for a console app. That did the job and got to the next part.

3 hour technical interview, ~2 hours of HR talk & getting to know me, then a set of 5 algorithms tasks. I failed at some point, as i'm not fan of algorithms tasks and i don't have CS education, so my algorithms knowledge isn't strong.

Result : Not qualified enough no offer

Company 2:

1 hour HR interview

After which they sent me a task that would have taken me about 8-10 hours to do. I politely Declined.

Company 3: 1 HR interview 2 x 1hr technical interviews,

First technical interview : extend a very very simple code (2-3 classes) with couple of functionalities and write tests. All in all it took an hour and was pretty straightforward. With no hidden gotchas

Second technical interview : you are given task details, and have 20 minutes to come up with overall system design. after which we spend about 30 minutes going trough that and asking why i choose X or Y, or what if the requirements changed to X or Y.

then 1 hour final "culture fit" interview. Result: waiting for offer.

Company 4: 2 x 1 hour interviews with the CTO & CEO. Tech talk mostly, but no deep algo questions. What challenges I've had and how I've solved them. Result : waiting for offer.

Company 5: one 1 hour HR interview, and i have to do a 5 hours long fixed time (i.e. they sent tasks at 12:00 i have to be done by 5 and send results).

And i'm debating if i should do it at all.

I've had the most pleasant experience with companies 3 & 4. From their HR staff to their Interviewers.

About about me for context: I'm a backend/full stack engineer with about 10 years of experience. I've coded professionally in PHP, C# and JavaScript. I've led few interviews, but never used spend-a-day-doing-it code challenges.

Company 3 is very close to the interview process that led me to my current job. They also gave me a choice between a "waterfall" approach, or pair programming where we talk about the solution as we go.

Most pleasant interview experience I've ever had.

Don't know what to say, but I think this is one way to iron out senior folks who don't do daily coding and most of the time was on designing, diagnosis or debugging.

When you go to interview,you may be interviewed by a fresh graduate who throw you a trick leetcode question that is, pretty much, requiring you come up with code in 30 minutes. Without a few months careful practice, it's hard to get it right under stress.

Coding challenges are like driving, you drive at normal speed in real life so you can drive everyday safely, but for interview, your brain has to be speed up the max speed under pressure, not everyone, not even the brightest guy who can do this well, again without months preparation.

I got interviews at Amazon etc but I never had time to practice leetcode etc, there is no way I can pass its whiteboard process without at least 3 months preparation I feel, and I absolutely have no time for that.

How about hiring senior folks as a contractor for 1 or 2 months then to decide the level and pay and such? nobody wastes anytime and no obligation either.

I disagree with the author. A coding challenge is a good way to judge how a candidate approaches a problem. The ability to walk through what you’re thinking while you’re doing it is a great judge of ability.

It also helps to weed out people who can’t actually program, this comes up annoyingly often.

I agree that having intentional errors in a starter repository is bad though.

You can also have a discussion one how one would approach the problem vs actually doing the coding. The discussion alone would give great insight into how the applicant understands the whole concept, and any potential pitfalls that can be anticipated, etc.

The prevailing opinion in this thread is that coding challenges are fair and equitable. HN was singing a different tune on the same subject just a few months ago. What changed? The economy collapsed. The labor market is no longer in favor of those selling their time and energy. So perceptions of this practice have changed.

Instead of a data structure and algorithms problem, give someone a broken webpack and/or babel configs and 30 minutes to figure it out. Will let you know if they understand bundlers and transpilers, which is probably more useful than an esoteric tree traversal question, at least with respect to front-end engineering.

Ugh, so many points I would like to address, but can't without turning this into a lengthy essay.

Anyway I'm currently interviewing and have been on both sides of the recruitment process.

I don't mind homework, but I know that more often than not it's never read or taken into account.

I discovered this by accident when I published my solution through a service that sends a notification when somebody downloads. Eventually they did, but only after I passed all the other steps - even though the task was given early in the process.

I've been publishing them like this ever since and this is my experience so far.

So yeah, coding challenges apparently are just a way to gauge your level of commitment, not a way to test your skills.

Personally I'm turning this around by asking to see a piece of code before we discuss compensation in detail. Dodged at least one bullet this way.

The fact is, that without some demonstration of a candidate’s usefulness (their code) I cannot hire them. We used to use HR, now it is the head of engineering who runs our technical hiring. The process is very simple, complete a task, and if it is done correctly we start looking at soft skills. If you can’t do it in a manner that suggests you can hit the ground running, you are cut.

Reality is harsh, and unpleasant. In an environment where resources are tightened by rapid growth, we cannot afford the trappings of these supposedly “superior and modern” hiring techniques. We need talent, not mediocrity.

I've been on both sides of the table and I'll tell you a coding challenge is a much better solution than the whiteboard/leetcode/code-on-the-fly during the interview. Just keep it short i.e. 2-4 hours max.

As far as alternatives to a coding challenge, the OP did suggest walking through a prior project on GitHub. I think that's a great idea (in addition to a coding challenge though); an open source project on GitHub is a bonus for a candidate, and will in many cases steer the interviewer to ask questions about your project in lieu of abstract concepts.

It's a tough problem, but eventually you have to demonstrate some sort of competency. White board, open source contributions, homework. There has to be some way of demonstrating you know what you are doing.

I'm a consultant so I've been through tons of interviews at this point in my career (21+ years).

If I had my choice? I'd pair program with my potential teammates on a bug/feature for the product I'd be working on.

Give me something real that's been in the backlog for a while, and we'll pair on it for the allotted time.

This sounds good, but it will not work for many jobs. Fixing a bug in a code base that you don't recognize is more difficult that a well defined coding problem. Candidates will be limited to the technologies and problems which they are already familiar.

This approach has some value, but it can't be the only way to evaluate candidates.

Yea, it's meant to be a waste of your time.

More accurately, if you need to learn new things to complete it - it will be very time consuming.

Someone experienced, or that uses relevant tech/skills on a daily basis should complete it in less than the time allotted.

Therefore "homework" coding challenges are useful for discerning if a candidate has relevant, current experience.

See the article on college professor solving a 30+ of CS problems in one day.

Whiteboard challenges, on the other hand, evaluate a candidates thinking process & presentation skills.

Whiteboard challenges

I don't like coding challenges, but on the hiring side it turns out even really basic ones can filter out a huge chunk of people who can't code their way out of a paper bag. But for most jobs they don't really need to be anything crazy. It turns out that plenty of applicants have a hard time with things that are surprisingly simple and that's really all you need.

I got chance to appear in a couple of coding challenges while applied for jobs. In one test I failed. It was a typical hacker rank riddle. The other I passed as I see my test ran properly yet I got a rejection letter. Another HN based job interview I was asked for a challenge but I simply refused.

Recruiting is mostly about finding reasons to say no.

And nothing gives more open ended reasons than the coding test.

Crossing concerns here:

Assignments can be helpful, but they can also be time consuming which is unfair.

Do the code or whiteboard 'on the day' - enough with the take home tests, or, pay people to do the tests.

What other industries do something similar in their hiring process?

A family member is a chef, and for new jobs you have to "stage" [0] where you come and cook for a (unpaid) shift to see if you are a fit, have the skills, etc. To me this is exactly the same thing, and I expect this is very common across different kinds of jobs.

[0] https://en.wikipedia.org/wiki/Stage_(cooking)

One thing I always wonder, how do other professions do it? Do civil engineers have to make a building? Do chemists have to assay a sample? What makes programming so unique?

Well for one you wont see Chemists or Civil Engineers from the variety of backgrounds developers come from. It's near impossible to become one without the degree.

I've been through so many puzzle dungeons that I immediately just exit a process if it involves another technical interview. I've done technical interviews with FAANG and the like and those processes actually tend to be much nicer at more mature companies because they're actually well designed problems with people who actually understand their significance. You at least get to ask questions and not just get told to jump through the hoop like a dolphin. For the rest, you end up with interviewers who google some esoteric Python problem they think is particularly clever.

I'll stop here to state that I think any adherence to being overtly clever is my first and only red flag.

It's so pathetically obvious because the number of times I've done the exact same problem for multiple companies because it happens to be the 2nd or 3rd search result.

I once had someone disqualify my answer for using a DFA because it wasn't "real code".

To be honest, I think the technical interview actually IS important, and I've had numerous experiences enjoying it if only to showcase a lot of the more necessarily toxic behaviors of a company up front. Let's be honest with ourselves, is any employer going to necessarily have the best practices with respecting your time, money and effort? It's best to lay out those cards up front. Literally every red flag will demonstrate itself through the interviewer in most scenarios. Like any test, you get out of it what you give.

For example:

How many times do I feel this?

"We take an extra time-slot / day / email chain in what is already a hypercompetitive market just to suss you out. We wait until the last stage of the interview process to do this because we still don't totally trust you or understand anything you've done. We waste a lot of time creating elaborate puzzles that don't really matter. We're not going to pay you because our finances are already horribly mismanaged that it's complicated to contract you out for the day and also we feel like you should be gladly externalizing your labor costs for our shiny position if you really love us. Drink the kool-aid."

Sometimes the technical interview surprises me. I always use it as a motivating example of myself as a candidates to work through a problem, show grit and also demonstrate when to ask for help. Ego and pride on both sides of the equation come into play and that's always a worthwhile exercise before engaging it any longterm commitment. A good test accommodates and informs through failure. I learned this from administering the Putnam. Never got a decent score on it, but it was a worthwhile experience to fail. Interview tests are on the other hand, are not designed to teach anything except that you are unworthy.

My best roles have never asked for it. The company knew what they were looking for, were willing to trust me, were able to move along the interview process without losing me to a competitor and everyone was happy and they always pay competitively. Usually if there's a technical interview involved at all, they have no idea what I can actually do. If they're willing to waste time, effort and money on extra steps and are taking "necessary precautions" to subject you to what are essentially just litmus tests, then I'm sorry, that speaks to me a sense of your culture of mediocrity and dogmatic orthodoxy towards process.

Goodhart's law. When a measure becomes a target, it stops being a good measure. When I was in school, we had other students take a year off to study for the Google interview. These were all affluent white men. At the same time, I had recruiters complaining to me that there were no decent black/brown candidates in the pipeline. I remember hearing the same thing except as a child. Uhhh, there's been plenty of time to invest into the local community. Recruiters would do well do understand to the history of the SAT as systemic racism, or literacy tests as systemic racism, or like any actual statistics as systemic racism. Everything that reduces candidates to numbers and quizzes feels rife for abuse. Who has the privilege to study an extra game on top of performing in a society that is already stacked against them? That's not a question.

To wit, there's the issue with a large number of candidates with junk resumes. There's the outstanding issue of "authenticity" in professional acumen. There's the whole problem with the industries underlying predatory behavior with regards to labor externalization and laundering numerical metrics, be it the rate of inclusion and diversity or the amount of hours actually demanded by management. What is so fundamentally wrong with the technical interview is that it's just the cherry on top for what is already a miserable process. I'd so much rather just do free work for a potential client and just give it to them so it actually means something and is useful than waste any more time solving pointless puzzles, and this is from someone who specifically enjoyed studying hard pointless puzzles in school.

No, you can't just show side projects. You should not be showing your work portfolio to anyone who isn't paying you. If they were interested in open source, they'd already be paying you.

The technical interview is dead. Long live the technical interview.

Most of the alternate approaches have their own severe failings:

Ask pertinent questions - a lot of candidates are bad at immediate summation and dissemination of information. There are many good engineers who have not learned how to sell their work, themselves, or in general communicate to the intent of a question. This is an invaluable skill, largely orthogonal to writing code, but it's a skill that's often learned with experience. Learning to ask level appropriate questions is hard and evaluating answers consistently (if not objectively) even harder. It is also largely insufficient in gauging programming ability. I've had co-workers who can talk about advanced architecture-y topics seemingly competently but struggle to write even relatively simple code.

Probationary period - I've never seen this implemented (excluding internships which sort of fall into this category) and it certainly sounds like a nice option to have available but I would bet many candidates would be afraid of a short probationary period (say 3 mo) especially if they have to move their family for the job. Being a professional and being told you are on a probationary period is stressful. At many large companies it can easily take more than 6wks to get up to any sort of productivity.

Walk through existing projects - most people (vast majority) do not have any sizable or clean or even "interesting" side project they would want to walk an interviewer through. Unless you worked on open-source for your previous employer (again minority of people) or on your spare time, this is not something you can get a strong signal out of.

Ask them to build something - err, I don't see how this is that different than a coding challenge. And again, not everyone has personal projects ready to be shown to an interviewer. For example, the game I'm coding some nights after work does not look anything like code I write for my employer; there are no tests, there's a lot of experimental and often dead code, functions are not neatly factored and thought out with an eye for collaboration, etc. I would not want to be judged by this code for an interview.

Pair programming - this sounds like a coding challenge but with a more empathetic interviewer. I don't think this is an alternative, just a mild augmentation of a coding challenge.

Focus on ability to learn - companies do this but it's usually considered supplementary to coding ability. Some companies explain their product and try to see if the candidate understands where the value of the product exists or at least where some of the challenges are. This could be spun out into a product type interview or a system design interview depending on intent.

Alternatives to white-boarding or do-at-home project interviews: - Provide a small existing but functional code base and ask the candidate to implement additional functionality. Downside is this takes a lot more effort to prepare and are limiting the candidate in regards to language they want to use. - Provide multiple options for a coding project or question and let the candidate choose which one they want to solve. This would hopefully reduce some of the luck factor of getting a question you know how to solve. - I can't think of any others, but I'm sure they exist

One of the big desirable goals of interviews are evaluating candidates in a consistent manner. All interviewers have biases and these biases vary between the interviewers. Trying to account for that is, I believe, important but also very hard. Using the proposed alternatives would make it even much harder.

so much complaining

I usually tell interviewers I don't do per-interview assignments. I'll often e-mail them a list of current open source taks I'm working on, a paragraph with good requirements or a linked issue from an issue tracker. I'll list the programing language and framework and ask them to select from the list. I'll complete my task, with tests, and do a full presentation.

Universally I get turned down. "You have to do our assignment." Oh and I'll usually offer to do their assignment for $300. That doesn't go well either.

Only one company took me up on this offer, and it's the one I work for. (They didn't even ask me to complete one of my tasks, they just looked at the project I authored and said that was fine; and scheduled a full interview).

The few times I have actually done these assignments, I've never gotten the job.

Their loss.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact