AFAIK (as someone who does not work at Google), "brain teaser" questions have not been asked at Google for almost a decade.
Every time I see complaints like this I can't help but think the posters have a chip on their shoulder from being rejected by Google or a similarly selective company. It's never a commentary on any fundamental issues with these types of interview questions and no one ever comes up with a better process that is equally scaleable and effective for hiring generalists. In my experience interviewing with dozens of small, medium and large companies, the vast majority of technical roles require these types of questions now. These come off as criticisms against Google specifically for asking hard variants of these questions.
I don't personally have any issue with companies asking questions like these as long as they don't simply look for "a correct and optimal solution coded up perfectly whilst under stress in under 30 minutes", but rather the process of solving the problem.
If we're talking about FAANG, I don't believe that they really care about "the process of solving the problem". If you don't come up with a close-to-optimal solution in about 5 minutes, you're in trouble - there will be competitors who are better prepared and thus can. Solving algorithmic questions is a very trainable skill, after all.
Funnily enough, this response is exactly what the parent seems to be talking about. FAANG originated as a stock term and not as a grouping of companies with similar hiring processes. i.e. Interviews at Netflix are generally domain specific. Likewise, Amazon isn't the same as Apple isn't the same as Google.
> If you don't come up with a close-to-optimal solution in about 5 minutes
Disingenuously hyperbolic. I was only given a single problem to solve in each of my 45 minute "coding rounds" at Google and Facebook. If you approach these interviews like a competitive coding competition you will absolutely not get the job. By that I mean arriving at a non-optimal solution as fast as possible while writing messy code and not explaining your thought process. Not to mention that senior roles involve an increasing number of interviews that focus on design rather than algorithmic questions.
This all seems like nonsense coming from individuals who are (understandably or not) frustrated at an interview process that doesn't align with their strengths.
But their hiring practices are pretty similar (perhaps except Netflix, don't know anything about them).
In one of my interview rounds at google I was under huge time pressure, despite coming up with an optimal solution in seconds. That was an outlier as the interviewer was not aware of this solution, so explaining it took a long time. But still.
The valuable part from the competitive coding background is not the coding, but "inventing" solutions. In quotes because it's mostly a pattern-matching process. And it's a _huge_ advantage.
Personally I'm perfectly happy with the current process, as I took about half a year to become good at it, and the payoff is nice. :) But I'm convinced that's what this process selects for first and foremost - amount of preparation.
Funnily enough, this response is exactly what someone who succeeded under this model would say.
This all seems like nonsense coming from individuals who are (understandably or not) frustrated that others would suggest that their success was only due to an interview process that aligned with their strengths, rather than their obviously superior skill and intellect.
I don't think I would be considered a "success" under this model as I never passed a tough interview and currently work at company you will never hear of. I simply have no desire to justify my skills by talking others down and pretending I'm too good for companies that reject me.
I'm not sure how this is a valid rebuttal to my comment since it addresses none of the point I made. Do you really think the secret sauce to getting into highly competitive jobs is by incoherently speed running through algorithm questions?
I just took your first and last sentences and had some fun with it.
That way of interviewing is good, because it weeds out people who don't prepare and thus might be lazy or just looking for an easy paycheck. But it also tends to confuse good memory of technical and abstract questions as competence, and if you're not careful you'll end up with a lot of copy pasted humans that are really good at inverting binary trees, but suck at everything else. Also people who have already passed it tend to hold on to it as pride, and might fail an otherwise amazing candidate, just because of a failure on a question that has no practical use outside of a textbook. Those are my thoughts on it, and I'd wager the thoughts of many others who are critical of the model.
> Do you really think the secret sauce to getting into highly competitive jobs is by incoherently speed running through algorithm questions?
No, the real secret sauce is having a friend or relative on the inside. If you don't have that it's a gamble most of the time imo.
Yes. Thank you for explaining my comment to me, and you're completely right, but you missed the part where my comment was based on his first and last sentences only going the other way around. It did seem a bit dismissive and condescending at the time so I had some fun with it.
And the only reply that to me somewhat seems to understand what's going on, getting lots of downvotes, so it got collapsed and you cannot easily find it:
> They just play a numbers game, and are able to discard perfectly capable candidates with this kind of questions, and still get a bunch of great candidates
Every time I see complaints like this I can't help but think the posters have a chip on their shoulder from being rejected by Google or a similarly selective company. It's never a commentary on any fundamental issues with these types of interview questions and no one ever comes up with a better process that is equally scaleable and effective for hiring generalists. In my experience interviewing with dozens of small, medium and large companies, the vast majority of technical roles require these types of questions now. These come off as criticisms against Google specifically for asking hard variants of these questions.
I don't personally have any issue with companies asking questions like these as long as they don't simply look for "a correct and optimal solution coded up perfectly whilst under stress in under 30 minutes", but rather the process of solving the problem.