6 ms·

Every time one of these interviewing posts bubbles up I skim it to see if the author mentions things like: company size, team composition, the nature of the wor

Every time one of these interviewing posts bubbles up I skim it to see if the author mentions things like: company size, team composition, the nature of the work the team is doing, the nature of the industry the company exists in, the way hiring decisions are made, the desired properties of their hiring process, their offer rate, acceptance rate, turnover rate, the amount of time positions tend to stay open for, or really just anything that would offer some context on what, specifically, their interview process is optimized for or achieving. Nine times out of ten that stuff is absent and the post is just a bunch of opinion and conjecture. By the way, when you're doing something nebulous in an interview like seeing "how the candidate thinks" what you're often doing (all too frequently without realizing) is looking for justification for a decision you've already made based on whether or not you like the candidate. The fact that you have them writing code on a whiteboard doesn't instantly make the whole thing an objective process. EDIT: You can test the above by making a "prediction" about the candidate's performance after the intro and comparing your prediction to your final decision after the interview.

anonymoushn 6y ago
Re: EDIT: It seems like you might accidentally be testing your ability to predict their ability to solve the whiteboarding problems given their resume. As an example, at one company we noticed that applicants with a certain level of educational achievement were dramatically less likely to be able to solve the problems we used than candidates with slightly higher or slightly lower levels of academic achievement. So if I made predictions about candidates, that's one factor that would influence my prediction, but it wouldn't necessarily mean that my subconscious bias was causing the candidates not to come up with a BFS.
- aahortwwy 6y ago
  I could've phrased that better. You're making a prediction about your ultimate hiring recommendation, not about whether they complete the problem or answer your questions correctly or not. The point of this exercise is not to produce useful data but to draw your attention to your own decision-making. > not to come up with a BFS But it's not about whether they solve the problem, it's about observing their problem-solving process... right? ;)
  - anonymoushn 6y ago
    Sure. At some other company, I interviewed a bunch of candidates in a section that was mostly about knowing how the protocols we dealt with work and how one might build a toy version of the software we worked on. For a section like this, it's easy to help candidates too much or too little, or to let one's preconceived notions about the candidate color one's recommendation too much. I am again concerned that I may have some actual ability to predict the candidate's performance (to the extent that people's knowledge of protocols and understanding of how to write server software exists as a real thing rather than a thing that I made up) from their resume, but now I additionally worry that as far as debiasing goes, just making predictions and comparing them to outcomes will not be doing nearly enough. I would prefer not to have such a section, but I am usually not in charge of these things. > But it's not about whether they solve the problem, it's about observing their problem-solving process... right? ;) I did not say that the interview process at that company was my favorite or was the same as the one in the OP. That said, my experience as a candidate at FAANG and smaller-but-not-super-small companies has generally been that I must at a minimum solve the problem, and that sometimes I must also solve it exactly the way some rubric says it is to be solved, so that providing an unusual approach that uses half the space and just as much time as the best approach the department has heard of also causes one to fail.
mpweiher 6y ago
> decision you've already made Abso-fucking-lutely, pardon my French. To be a little more precise: unless you're going into an interview totally unprepared, there is a résumé that you've read, hopefully read carefully. I make a decision, based on said résumé, whether I would like to hire that candidate. Maybe "decision" is too strong a word. Hypothesis. Yes, hypothesis is better. That means I have some sort of mental model of what the candidate might be like, what they can do and whether and how they fit. If the person is invited for the interview, the hypothesis is "yes, I would like to hire this person". The purpose of the interview is to try to disprove the hypothesis. If you fail to disprove the hypothesis, hire the person.
- aahortwwy 6y ago
  A lot of companies instruct their interviewers to "default to no" when making hiring recommendations. Effectively the opposite of what you're suggesting. I knew a lot of interviewers at those types of companies who wouldn't bother to read resumes, either. They'd cite "bias" or something.
  - reallydontask 6y ago
    I recently applied to the Civil Service in the UK and they asked me to produce an anonymised CV, namely remove my name and contact details In principle I like the idea but I think they should ask to also remove the company names and perhaps dates, as it could be straight forward to work out who you are based on companies and dates. As to how effective to remove bias, it's hard to say as I've not seen any A/B testing done with this, but give the massive push to get minorities on board, I'd hazard a guess that it's not as effective as they had hoped
  - mpweiher 6y ago
    > "default to no" .. opposite of what you're suggesting. Not necessarily, just depends on how easy/hard you set the bar for disproving the hypothesis. It also is somewhat inherent in that you really are trying to disprove the hypothesis. On the other hand, if, after reading the résumé, I do not want to hire this person, why on earth should I go on to interview said person? The "default to no" part is, IMHO generally correct, because false positives are a lot more damaging/costly than false negatives. My perspective may be warped here, because I have worked for companies that get LOTS and LOTS of applicants. At some the team would screen hundreds of résumés a week, of which maybe 1-2 were somewhat interesting, and that was after the HR screen. Oh, and I also got to clean up after one of the candidates → colleagues I said no to and the rest of the team approved... ¯\_(ツ)_/¯

anonymoushn 6y ago

Re: EDIT: It seems like you might accidentally be testing your ability to predict their ability to solve the whiteboarding problems given their resume. As an example, at one company we noticed that applicants with a certain level of educational achievement were dramatically less likely to be able to solve the problems we used than candidates with slightly higher or slightly lower levels of academic achievement. So if I made predictions about candidates, that's one factor that would influence my prediction, but it wouldn't necessarily mean that my subconscious bias was causing the candidates not to come up with a BFS.

aahortwwy 6y ago
I could've phrased that better. You're making a prediction about your ultimate hiring recommendation, not about whether they complete the problem or answer your questions correctly or not. The point of this exercise is not to produce useful data but to draw your attention to your own decision-making. > not to come up with a BFS But it's not about whether they solve the problem, it's about observing their problem-solving process... right? ;)
- anonymoushn 6y ago
  Sure. At some other company, I interviewed a bunch of candidates in a section that was mostly about knowing how the protocols we dealt with work and how one might build a toy version of the software we worked on. For a section like this, it's easy to help candidates too much or too little, or to let one's preconceived notions about the candidate color one's recommendation too much. I am again concerned that I may have some actual ability to predict the candidate's performance (to the extent that people's knowledge of protocols and understanding of how to write server software exists as a real thing rather than a thing that I made up) from their resume, but now I additionally worry that as far as debiasing goes, just making predictions and comparing them to outcomes will not be doing nearly enough. I would prefer not to have such a section, but I am usually not in charge of these things. > But it's not about whether they solve the problem, it's about observing their problem-solving process... right? ;) I did not say that the interview process at that company was my favorite or was the same as the one in the OP. That said, my experience as a candidate at FAANG and smaller-but-not-super-small companies has generally been that I must at a minimum solve the problem, and that sometimes I must also solve it exactly the way some rubric says it is to be solved, so that providing an unusual approach that uses half the space and just as much time as the best approach the department has heard of also causes one to fail.

mpweiher 6y ago

> decision you've already made Abso-fucking-lutely, pardon my French. To be a little more precise: unless you're going into an interview totally unprepared, there is a résumé that you've read, hopefully read carefully. I make a decision, based on said résumé, whether I would like to hire that candidate. Maybe "decision" is too strong a word. Hypothesis. Yes, hypothesis is better. That means I have some sort of mental model of what the candidate might be like, what they can do and whether and how they fit. If the person is invited for the interview, the hypothesis is "yes, I would like to hire this person". The purpose of the interview is to try to disprove the hypothesis. If you fail to disprove the hypothesis, hire the person.

aahortwwy 6y ago
A lot of companies instruct their interviewers to "default to no" when making hiring recommendations. Effectively the opposite of what you're suggesting. I knew a lot of interviewers at those types of companies who wouldn't bother to read resumes, either. They'd cite "bias" or something.
- reallydontask 6y ago
  I recently applied to the Civil Service in the UK and they asked me to produce an anonymised CV, namely remove my name and contact details In principle I like the idea but I think they should ask to also remove the company names and perhaps dates, as it could be straight forward to work out who you are based on companies and dates. As to how effective to remove bias, it's hard to say as I've not seen any A/B testing done with this, but give the massive push to get minorities on board, I'd hazard a guess that it's not as effective as they had hoped
- mpweiher 6y ago
  > "default to no" .. opposite of what you're suggesting. Not necessarily, just depends on how easy/hard you set the bar for disproving the hypothesis. It also is somewhat inherent in that you really are trying to disprove the hypothesis. On the other hand, if, after reading the résumé, I do not want to hire this person, why on earth should I go on to interview said person? The "default to no" part is, IMHO generally correct, because false positives are a lot more damaging/costly than false negatives. My perspective may be warped here, because I have worked for companies that get LOTS and LOTS of applicants. At some the team would screen hundreds of résumés a week, of which maybe 1-2 were somewhat interesting, and that was after the HR screen. Oh, and I also got to clean up after one of the candidates → colleagues I said no to and the rest of the team approved... ¯\_(ツ)_/¯