[jknz]

Criticism of interview practices is very common. However, I can't imagine that large companies that hire in the hundreds/1000s every year (say, FANG) don't have enough data to have optimized their interview practices.

These companies know in the last 10-20 years who interviewed who and the performance of everyone after hiring. With enough data you can:

- A/B test which interview questions better predict future performance

- how much each interviewer's feedback is correlated with future performance of the interviewee. Some interviewer's are likely terrible and they should not interview. Some interviewers at good at predicting future performance and they should interview much more if not full time.

- which interviewer provides biased feedback (women, minorities, etc)

Either these companies are not analysing this data and the interview practices may indeed be garbage. Or FAANG companies are analysing this data and otpimizing their interview practices accordingly and the process/results are actually as fair and good as it can be.

[username90]

Laszlo Bock found that interview skill is pretty homogeneous, so it doesn't really matter who does it as long as you just do enough of them. But if you do 4 algorithm interviews you have pretty good interviews.

> Four meticulously orchestrated Google interviews could identify successful hires with 86 percent confidence, and nobody at the company—no matter how long they had been at the company or how many candidates they had interviewed—could do any better than the aggregated wisdom of four interviewers.

https://www.theatlantic.com/business/archive/2016/04/the-sci...

[naniwaduni]

Or perhaps incentives are misaligned. (When are they ever not?)