Hacker News Reader

>There's a decent amount that you can do to prepare for one of these things, but there's also an incredible amount you don't have control over.

This is super important. Don't take it personally at all. They're using the randomness as a flow control. They are always attracting more talent than they actually need purposely to keep the reservoir full and just running the extra through the spill gates.

Just taking a ride over the chute is proof positive that you've got marketable talent.

> If you don't get the job, it's not a signal regarding your quality as an engineer and you shouldn't interpret it as such (however tempting it may be to do so).

Of course "not getting a job" is a "signal regarding your quality as an engineer". We need to keep this toxic mindset out of tech and silicon valley. The idea that everyone is smart, special, and deserves a job at Google, is completely false. There are people here on HN that will never get a "top" software/machine learning/design job simply because they are bad at it. Telling them "it's not you, it's the company" will only doom their future even further as they will lose motivation to improve themselves because: "Hey, HN said I'm smart and a talented engineer."

To drive the point home to viewers reading this, some of you may have applied to Google via their website. You did not get a reply. After reading this original post and the GP, please do tell me how smart you are and why your 10x programmer skills didn't catch the eye of Google recruiters (or Facebook, Apple, or any hot startup).

I believe what he was trying to say is that there is a large factor of chance involved with interviews.

Therefore it should not be a signal. Not getting a job because of an interview does not mean that you are not a qualified engineer.

It may mean that though. But it's not as if it's a 100m sprint where you can very easily measure results between Olympian and couch potato.

> https://www.amazon.com/The-Drunkards-Walk-Randomness-Rules/d...

There is a large factor of chance if you're a marginal candidate. But if you're very bad, there's no chance, and if you're very good, there's a very high chance.

Good at interviews? Then yes...

On the other hand some of us know we aren't top class engineers and don't actually care, I'm content with been good and careful over been 'Google' class.

Not that I'd want to work for Google anyway, I don't like them very much as a company.

I failed my first interview and managed to get it on my second and so did a number of engineers at Google, Facebook .. you name it. The point is these interviews are intended for zero false positive and are very likely to result in false negatives. "Not a getting a job" alone doesn't state anything about your quality as an engineer and regarding catching recruiters eye, I know a ton of people who didn't get any responses for job applications in college, but a year or two at a good role made recruiters flock to them. My two cents would be there is always room for growth in software engineering and your work is the only signal that you should rely on. (coworkers, open source contributions etc). I wish there is a way to evaluate people solely based on their but unfortunately that utopia is far.

I was rejected from Google when I first graduated from college. Now I work at NASA and I'm very thankful I didn't get that job.

Saying that the hiring process is flawed isn't the same thing as saying everyone is special and deserves a job there.

From my anecdotal experience, I felt that they focus too much on algorithms and give the same interview to every candidate regardless of their experience or desired job. I made it very clear I wanted to work in embedded systems and write code for microcontrollers and they asked me questions about graph theory. Algorithms are not everything, especially for a company with such diversified technology opportunities. If I'm working on search, then sure. But embedded software? Can't they take the time to focus the interview even the slightest bit?

Not a big deal though. I know they have an endless supply of candidates and thankfully we as software engineers have a pretty good selection of places to work.

> I made it very clear I wanted to work in embedded systems and write code for microcontrollers and they asked me questions about graph theory. Algorithms are not everything, especially for a company with such diversified technology opportunities. If I'm working on search, then sure. But embedded software? Can't they take the time to focus the interview even the slightest bit?

I'm not a Googler, but from what I hear, moving between teams within Google takes less effort than most places, so that might explain why Google tends to interview for a generic "Google developer" rather than a "Embedded developer" position. Who knows, you might progress from initially working on the Google home firmware, then move on to optimizing the search client code and then possibly the search backend for Home itself.

I'm sure there are a lot of specialists at Google, but I suspect those are fewer than generalists with a solid handle on data structures and algorithms.

They are starting to take into account area of expertise, like in the article, they offered 2 Android-specific interviews as a different process. You'd still have some algorithms, but at least the other 2 would be for embedded if that was your area.

I think what he was trying to say is that it's not an binary relationship.

I used to work with someone who now works at Google. He was the sole lead for a project that he absolutely butchered the code before bailing out. I was left to clean up the mess which took many months and then continuing to finish the project to completion to the satisfaction of the end customer. I can certainly say he was not 10x programmer but perhaps he found some niche role there that he can contribute towards.

you missed the context. OP was describing ways in which qualified candidates are rejected anyway because, in that particular time and place, there were even better candidates also interviewing, or that the teams that were hiring changed their priorities, or whatever.

What a sad world we live in where you define your "quality" as your worthiness to serve someone else. So pitiable.

One of the most frustrating things I've seen is that interviewers/recruiters themselves couldn't stand up to the same rigor of questions which they expect from applicants. In one such company, I asked a question on "algorithmic fairness, accountability and transparency" about core algorithms they used for their business. First, it came across as a bit of a rude shock to them - they hadn't heard about it. Second, they parroted a line which was along the lines of "our legal team looks into it and makes sure it is legal". Third, when pushed about the fact that my question was not on legality - they refused to accept that they may have overlooked this part or were not aware of it. Rather than engage in a discussion, when the boot was on the other foot - they wanted to weasel their way out of it.

I've worked for a company that's given pretty much the same two hour on-site test for years, and several times the people in charge of the hiring decisions have forgotten how to do parts of it (it's an open office, I've heard them admit it to the candidates).

It amazes me how they rejected so many people on this test over the years (seems to be over 90% rejections), when they can't keep it straight themselves. Granted they were always busy on other things, but still. Even I barely passed the test.

That being said, they left you alone for two hours while you work on it, sitting in the same environment where other devs are working, which I found much better than some other interviews, the worst of which was coding a full project on a laptop for a platform I haven't worked on professionally in over a year and was a bit rusty on, while one person stared at me from across a desk and a second person stood behind me and looked down at the screen over my shoulder.

I wonder if google had to reinterview their existing employees, in a blind study...

How many rounds of interviews before they eliminated most of their staff?

Once upon a time, a hiring committee was presented some feedbacks and had to come to a yes/no decision for some candidates. All them were rejected. It turned out that the feedbacks were of the hiring committee's members, from when they got hired.

Source?

It was in this video but don't remember exactly when. https://youtu.be/r8RxkpUvxK0

Direct Link: https://youtu.be/r8RxkpUvxK0?t=530

Yup, this is the video!! Not an urban legend.

I've heard this before, is there a source for this? It has all the hallmarks of an urban legend...

I don't have the source anymore but it was an article where they spoke to someone at Google HR who mentioned that a particular team was difficult to approve anyone or hiring. To prove their point, they anonymized the team's own resume and asked them to rate them.

I saw a lecture (about tech interviews) by a former Google engineer who at the time was working at Etsy. He mentioned that the experiment was performed on him and his team. The hiring committe he was a part of had acquired a reputation of being particularly difficult to please. They ended up rejecting all of their own hiring packets. I was skeptical when I heard the story. Google must do a fantastic job anonymizing the hiring packets that get reviewed, or maybe they had been at Google for a long enough time that they forgot their own process and didn't clue in. However, this engineer did mention it so it doesn't seem to be an urban legend to me. With some clever searching I'm sure you can find the video.

This is good if you assume that there's a desire to improve the quality bar over time.

You are making several assumptions:

1. That we can clearly and precisely define intelligence 2. That successful interviewing correlates with intelligence 3. That hiring people who can clear this gauntlet will generate more value for the company in the form of better products 4. That there is only one "bar" for quality

As far as I am aware, current research indicates intelligence is multi-variable and by extension there are multiple "bars" for quality one could measure. It also seems likely a team will be more successful if it includes people who are capable across multiple dimensions.

We're still in the very early days of Software Engineering. Why do we insist that the Landscape Architect must be familiar with the intricate details of HVAC Engineering?

Maybe they saw a bunch of people with outdated skills?

> There's a decent amount that you can do to prepare for one of these things, but there's also an incredible amount you don't have control over.

This is where I call bullshit. This person prepared for six months. You can't tell someone to not take something personally if he or she took six months of their life to study up, only to hit a dead end because your company didn't find it within themselves to take control of the things they could have.

I did the same and failed. I failed in a very specific way which annoyed me a little bit and which left me thinking how superficial this process is.

But 1. i extended my knowledge 2. i came closer to working for google than ever 3. if you really wanna work for google / <company> you will try again and even optimize those superficial 'issues'

I interviewed at Google with a no-hire. I asked for 3 weeks to brush up on algorithms and data structures. I should have asked for 3-6 months.

This is where it falls apart for me, though. Based on my experience, you really do need to be extremely sharp on data structures and algorithms, and you need to be able to do it at the whiteboard.

I have my own opinion on this, like everyone. Personally, I'm completely ok with requiring a data structures and algorithms "entrance exam" for many software development positions. I see it as similar to the vector calculus and linear algebra exam required of actuaries.

My problem is with the implementation. Actuaries take a well vetted exam, with a clear study path, and when they pass it, they get a lasting credential respected in their industry. In short, taking and passing this exam means they don't have to submit to potentially random and capricious (and unpredictable) exam based interviews (really, after 6 hours at the whiteboard doing programming problems why do we call this an interview? It's an exam, pure and simple).

Here's the thing - I'd do this for a lasting credential. I'd accept that this is a test I have to take, and properly study for. I'd take my 6 months, and I'd put in the time. What I'm not willing to do is take this test over and over and over, every time I interview at a new company. I just don't walk around with this stuff loaded into short term memory. I don't walk around ready to take my undergraduate data structures and algorithms test. And I think I've taken it for the last time.

Same for vector calc and linear algebra. I did well in these classes in college, and if I were interested in being an actuary, I'd study up and get back to that level I was at just before I took my exams in college. But no, I'm not interested in allowing a random person at a company to put me through the exams again.

This is why I'm pretty much done with technical interviewing and tests. I'll find some other way to participate in this field, or leave it altogether if that turns out not to be possible.

I do wish tech leaders, who so constantly bray about a shortage in our field, would recognize the extent to which their behavior and practices drive talented people out of programming and into other fields. They didn't stumble upon this shortage, they created it. And every crutch we give them just allows them to continue with business as usual.

I agree with the noise of the process. Actually, the whole experience was a good one for me, maybe I didn't express it in the post? Anyway, I got a job in Amazon UK at the end, so not too bad.

"not too bad"? Understatement of the year! ;) Congratulations on the job.

Thanks :)

I was unsuccessful in my interviews with Google several years ago, but left the process more impressed with the company and its employees than when I started. As a candidate, they were incredibly respectful to me, even when it became clear I wasn't the right fit for the roles they needed to fill.

I truly enjoyed them bringing me out for the weekend before my interview so that I wasn't rushed and to give me an opportunity to see the area. I also enjoyed the interview process itself (well, right up until the point one of my interviewers introduced me to some math I had no chance of solving). One of my interviewers even stayed in touch with me after the process, checking in occasionally and helping me make some new contacts.

Today, when I interview people, I keep my Google experience in mind and try to provide candidates a similar impression of my firm.

The randomness factor is a huge problem for Google and similar SV companies.

I'd be happy to take a chance on a Google interview. What I'm not going to do is take 3 or 4 days of my very limited vacation time to jump through Google's hoops, along with spending weeks of time reviewing obscure algorithms I haven't touched since college, and will probably never use again.

Honestly, I think if the H1B and other loopholes these companies use were tightened up, the interview process for engineers would quickly approach the respectful negotiations common with other professional industries (law, medicine, finance, etc).

I agree. After I have interviewed and seen Google pass on a candidate, I sometimes feel I was lucky in my interviews.

"If we didn't get lucky enough to hire you, someone else will quite soon."

The OP did get hired -- he mentioned that in his first paragraph.

"The OP did get hired -- he mentioned that in his first paragraph."

The OP was hired by Amazon, "we" refers to Google in this case.

I was referring to the"someone else" not the "we".

You think a company like Google would be able to implement a non-subjective and fair interview. Their current process must cost them a lot of talent

I doubt they are lacking good resumes. They can afford to miss good people

> Their current process must cost them a lot of talent reply

They optimize to reduce false positives. They're completely fine with false negatives. They don't have any shortage of talented engineers willing to interview there.

That last bit is why I think it's quite silly for smaller/unknown startups to be copying Google's interview process, as seems to be the trend.

Maybe it's part of the process. It's easier and cheaper to maintain good image to get people interested in working for Google (improving workforce supply) than to optimize the interviewing part.

It's the same with manufacturing. It's better to provide 96% reliability with warranty than to struggle with providing 99.9% reliability.

>>They optimize to reduce false positives. They're completely fine with false negatives.

Getting too many false negatives is a very sure way of getting a lot of false positives.

>>They don't have any shortage of talented engineers willing to interview there.

You might be wrong there. Apart from huge money opportunities, there are little incentives for any body to waste their best years in large company political middle levels.

>>Getting too many false negatives is a very sure way of getting a lot of false positives.

Why ? IMO if you reduce the False Acceptance Rate, you will up the False Reject Rate. Good chart at https://www.tractica.com/biometrics/in-biometrics-which-erro...

> Apart from huge money opportunities

It's...amusing that you brush that off like it's nothing.

Not to mention things like job security(people have families ya know?) and that having Google on your resume is still a powerful boost for future opportunities. That might not be true someday, but for now it is.

this is good analysis. Google does things that make sense for Google. unfortunately because they are prominent many companies that aren't Google copy Google's methods uncritically. it's a disaster. I can't tell you how many times I've cringed at small tech companies that run a glorified e-commerce website interviewing candidates as if their engineering staff is 1000 instead of the 8 that it actually is.

There's no such thing as a non-subjective (the word you're looking for here is "objective", I think) and fair interview. That's part of why it's so difficult. Everybody has a bad day--interviewers and interviewees. Some people can talk out a problem better than they can whiteboard, others the reverse. It is a wheelbarrow full of conflating factors and there's no magic bullet for it.

The tendency for engineering types to believe in "objective" processes for damned near anything that isn't measured with calipers is one of the bigger weaknesses in this field.

You also have to consider the cost of hiring the wrong candidate. That can often dwarf the loss from missing out on a good candidate. The other option is to hire less stringently but use trial periods which have their own problems.