Introduction
It's time to have a serious talk about interviews. It seems like everyone is talking about this topic, but we only discuss one side of the process — how candidates can successfully pass the selection. What about the interviewer? What makes them effective?
In recent years, candidates have become armed to the teeth: blogs, online courses, books, mock interviews, and methods like STAR have turned interview preparation into a full-fledged industry. Today's candidate knows how to approach system design, has honed algorithms, and memorized answers to typical questions. All else being equal, they will perform well.
But interviewer preparation often leaves much to be desired. At best, they watch a two-hour webinar on how not to violate labor laws and observe a couple of other people's interviews. Maybe they even conduct a few themselves. The result of this imbalance is a lot of poorly conducted interviews. It's time to change that.
Why interviewers make mistakes
Let's admit it right away: conducting an interview is a challenging task. The average interview lasts 45 minutes via video call. By the time you set up the equipment, chat about the weather, and answer the candidate's questions, you're left with at most half an hour for the actual assessment. And the candidate, as we've seen, has already seen all the standard questions.
In the past, you could stump a candidate with a tricky question, like how to delete a file named -f or sort a million numbers with minimal memory. That was a stress test for creativity. Now those questions have been thoroughly dissected. Hearing a perfect answer tells you only that the candidate prepared well for that specific question. No diagnostic of their thinking occurs.
So should we just invent new questions? That's a dead end. All problems that fit into 45 minutes have already been invented. New ones are just repackaging old algorithms. And even if you create a unique problem, it will become known to the industry after the first interview.
Solution: upskill the interviewers
My solution: improve the qualification of interviewers. If candidates evolve, interviewers must evolve along with them.
The main myth to dismantle: “A good interview = good questions.” No. A good question helps, but interviewer skill matters more. A professional will extract maximum signal from any question; a novice will miss every cue even with a perfect prompt.
The secret of a good interview is digging
Take the same question about a file named -f. The candidate answers brilliantly: rm -- -f. So what? You learned they memorized a trick. But if you ask:
Why does this work?
Would rm '-f' work? What about rm ./-f?
Does it depend on the shell or Unix version?
Would it delete a symlink or the file itself?
...then a real conversation begins. You see how the candidate reasons and applies knowledge of shells, syscalls, and utilities. And if they arrived without knowing the details but built a logical chain and proposed experiments — that is gold.
Training and motivating interviewers
The problem is that in many companies interviewing is treated as a chore to get through quickly. There is no training, no quality control, and — most importantly — no motivation to get better.
Companies do not need every engineer to interview. They can select people who want to grow in this direction and build an interviewing culture where craft is valued and rewarded.
Incentive systems matter a lot
“Show me the incentive and I will show you the outcome.”
— Charlie Munger
Money is not everything. Public recognition, a “Super Interviewer” mug, a directory badge, a mention at the all-hands — people need to feel the work matters. And interviewer skill should count toward promotion. Otherwise time goes into writing docs, not building an interviewer guild.
Improving interview quality is not that hard, but it takes time, money, and a systems approach. It is worth it if you are hiring the best.