4 ms·

I interview at Google and I never look at resumes - to avoid bias. I don't want to be biased by whether you went to Yale or UT Austin. My question is always a s

by pradn 8y ago

I interview at Google and I never look at resumes - to avoid bias. I don't want to be biased by whether you went to Yale or UT Austin. My question is always a standard algo/DS question anyway.

Cookingboy 8y ago
That is a good way to eliminate bias, but in the end, is that a good way to make sure each position is staffed by the best candidate possible? I always pondered about that when I was at Google. We had mostly smart people, but I argue not all of them were the perfect fit for the position/skills needed. Sure some bias are best eliminated such as what school they went to, but if work experience in general should not be discarded completely in my opinion and you can only get so much out of standard algo/DS questions, especially considering these days everyone is studying for those like standardized tests. I literally know some CS students who just did hundreds and hundreds of practice problems and they'd ace most big company algo interviews but that doesn't really tell me if they'd be good engineers in practice at all.
- QML 8y ago
  Sadly that’s the trade off when you want a common benchmark to measure students, but then students optimize for that benchmark instead of something else — basically Goodhart’s law. As a current student studying CS, I find all of this “leetcoding” disheartening since it unnecessarily takes away time and deprives some people of passion.
- aprdm 8y ago
  That reminds me of a conversation I had with a friend who works at Amazon in Seattle. He says they have to fire guys in his team frequently because they ace the interview process by practicing it ad nauseum but are terrible software engineers.
existencebox 8y ago
I say the following having worked at multiple "big-5" including G. Do you not feel any of the twist in your stomach that I feel when reading your post? That we're discarding any vestige of track record and accomplishment (that I pragmatically see as having correlated better than any other variable with effective coworkers in the past) for a readily gamable trivia question that I take some perverse pride in not having wasted time studying over the last decade? (Personally, I've spent my time doing far more actionable and applicable things such as running large distributed systems and learning how to do so more effectively while developing better patterns in that scope) To be clear, this is not some defense of bias. This is a statement that your stated approach seems to throw the baby out with the bathwater. If I think to myself of what this interview process incentivises, it's not behavior that would lead me to be the engineer I am today, nor the behavior that I would chose in a coworker. I'm just curious as to if there's any amount of dissonance inside G any more as to whether this is an effective way to get the people they want. Maybe it is, per their constant trumpeting of data driven hiring, and that's probably a great signal of why I didn't really feel at home there.
- sidibe 8y ago
  I can't speak for GP but I'm guessing they don't feel that taking the big picture view is their role in the interview process. In some companies the interviewer is just there to be another data point for the people who make the decision, and if their role is to just ask whiteboarding question it seems like looking at the resume could be the wrong thing to do.
- barry-cotter 8y ago
  > To be clear, this is not some defense of bias. This is a statement that your stated approach seems to throw the baby out with the bathwater. It is impossible to give every candidate an unbiased fair chance without throwing the baby out with the bathwater.
- sanedigital 8y ago
  > That we're discarding any vestige of track record and accomplishment... It's not your job as one of the five interviewers at Google to evaluate anything except the candidate's performance in your interview. The Hiring Committee (HC) then takes the feedback received from all the interviewers and combines it with the interviewee's résumé, any recommendations they've received, etc. Source: gave over 100 interviews at Google.
  - existencebox 8y ago
    Realize though, the unfortunate result of the way this is implemented results in an interview process that ends up being perceived as quite coarse and unfeeling, and seems to filter on often capricious tests of algorithmic knowledge; Some highlights in my personal experience, writing an RB tree, and "write a regex to capture the comments in code;" simply put I never felt that my experience was even slightly under consideration. I realize the typical response from a Googler (including my friends there now) is that these things specific instances are disallowed/not recommended, and I trust they're not lying with those statements, but for whatever reason _those did happen some years back_, and I continue to hear similar attestations both here and at work. Feel free to write off all of my frustration as a selfish fear that I won't be able to pass these gates in N years with the heavy focus on algorithms-under-pressure; it's certainly hit and miss now, and I'm open that I've been both accepted and rejected at all the places I've worked to call out how much of a crapshoot this process seems to be. I fully acknowledge that an org of G's size has needed to make certain tradeoffs/processes, but that has tradeoffs for candidates, which can be hair-pulling when we do/have done isomorphic work at a high level. (I should probably disclaimer that MSFTie, I have nothing particularly against G in this and all opinions are my own, but highly algorithmic interviews are a topic I've always been selfishly unsettled about)
deagle50 8y ago
Sounds like you're dismissing a person's accomplishments in favor of pseudo-science. What's next? Unconscious bias training and quotas? Oh wait...

by pradn 8y ago

I interview at Google and I never look at resumes - to avoid bias. I don't want to be biased by whether you went to Yale or UT Austin. My question is always a standard algo/DS question anyway.

Cookingboy 8y ago

That is a good way to eliminate bias, but in the end, is that a good way to make sure each position is staffed by the best candidate possible? I always pondered about that when I was at Google. We had mostly smart people, but I argue not all of them were the perfect fit for the position/skills needed. Sure some bias are best eliminated such as what school they went to, but if work experience in general should not be discarded completely in my opinion and you can only get so much out of standard algo/DS questions, especially considering these days everyone is studying for those like standardized tests. I literally know some CS students who just did hundreds and hundreds of practice problems and they'd ace most big company algo interviews but that doesn't really tell me if they'd be good engineers in practice at all.

QML 8y ago
Sadly that’s the trade off when you want a common benchmark to measure students, but then students optimize for that benchmark instead of something else — basically Goodhart’s law. As a current student studying CS, I find all of this “leetcoding” disheartening since it unnecessarily takes away time and deprives some people of passion.
aprdm 8y ago
That reminds me of a conversation I had with a friend who works at Amazon in Seattle. He says they have to fire guys in his team frequently because they ace the interview process by practicing it ad nauseum but are terrible software engineers.

existencebox 8y ago

I say the following having worked at multiple "big-5" including G. Do you not feel any of the twist in your stomach that I feel when reading your post? That we're discarding any vestige of track record and accomplishment (that I pragmatically see as having correlated better than any other variable with effective coworkers in the past) for a readily gamable trivia question that I take some perverse pride in not having wasted time studying over the last decade? (Personally, I've spent my time doing far more actionable and applicable things such as running large distributed systems and learning how to do so more effectively while developing better patterns in that scope) To be clear, this is not some defense of bias. This is a statement that your stated approach seems to throw the baby out with the bathwater. If I think to myself of what this interview process incentivises, it's not behavior that would lead me to be the engineer I am today, nor the behavior that I would chose in a coworker. I'm just curious as to if there's any amount of dissonance inside G any more as to whether this is an effective way to get the people they want. Maybe it is, per their constant trumpeting of data driven hiring, and that's probably a great signal of why I didn't really feel at home there.

sidibe 8y ago
I can't speak for GP but I'm guessing they don't feel that taking the big picture view is their role in the interview process. In some companies the interviewer is just there to be another data point for the people who make the decision, and if their role is to just ask whiteboarding question it seems like looking at the resume could be the wrong thing to do.
barry-cotter 8y ago
> To be clear, this is not some defense of bias. This is a statement that your stated approach seems to throw the baby out with the bathwater. It is impossible to give every candidate an unbiased fair chance without throwing the baby out with the bathwater.
sanedigital 8y ago
> That we're discarding any vestige of track record and accomplishment... It's not your job as one of the five interviewers at Google to evaluate anything except the candidate's performance in your interview. The Hiring Committee (HC) then takes the feedback received from all the interviewers and combines it with the interviewee's résumé, any recommendations they've received, etc. Source: gave over 100 interviews at Google.
- existencebox 8y ago
  Realize though, the unfortunate result of the way this is implemented results in an interview process that ends up being perceived as quite coarse and unfeeling, and seems to filter on often capricious tests of algorithmic knowledge; Some highlights in my personal experience, writing an RB tree, and "write a regex to capture the comments in code;" simply put I never felt that my experience was even slightly under consideration. I realize the typical response from a Googler (including my friends there now) is that these things specific instances are disallowed/not recommended, and I trust they're not lying with those statements, but for whatever reason _those did happen some years back_, and I continue to hear similar attestations both here and at work. Feel free to write off all of my frustration as a selfish fear that I won't be able to pass these gates in N years with the heavy focus on algorithms-under-pressure; it's certainly hit and miss now, and I'm open that I've been both accepted and rejected at all the places I've worked to call out how much of a crapshoot this process seems to be. I fully acknowledge that an org of G's size has needed to make certain tradeoffs/processes, but that has tradeoffs for candidates, which can be hair-pulling when we do/have done isomorphic work at a high level. (I should probably disclaimer that MSFTie, I have nothing particularly against G in this and all opinions are my own, but highly algorithmic interviews are a topic I've always been selfishly unsettled about)

deagle50 8y ago

Sounds like you're dismissing a person's accomplishments in favor of pseudo-science. What's next? Unconscious bias training and quotas? Oh wait...