Interviewing candidates who know more than you

Marco Caccin was a Machine Learning (ML) Engineer, who was the go-to choice for an interviewer for screening ML people for Lyft Europe.

Q: How did interviewing tech people for Lyft Europe go?

A: While at Lyft, I was doing a lot of screening interviews for European candidates, many of them for internship positions in machine learning teams.

Each candidate brings their own experience. My job as an interviewer is to tap into it. In many cases, that's a good return on the investment of the 2-4 hours spent for each interview (~1h for the proper interview, plus writing the report and an indefinite amount of time studying topics around the candidate's CV).

Q: What went easy?

A: Preparing for the coding challenge was relatively easy. The problems were picked from a pool of options. I had my favorite question with a couple of backups. Once I knew the space of solutions inside-out and had attended some interviews in "shadow mode", it was relatively straightforward to calibrate expectations of how a candidate should perform, while also figuring out if, how and when to give hints.

Why is it easy? Because this type of interview is rather quantitative and "mechanical": there's a playbook to follow, and it's hard to mess it up.

Q: What was your biggest struggle?

What I am really trying to achieve is to peek into someone else's thought process. When lucky, interviewer and interviewee are like-minded, and it all feels like an exciting tech conversation. Other times, I am left wondering - is the candidate shy, is there some cultural aspect I don't understand, or are they just not good enough yet?

Q: What did the interview process look like?

A: The whole interview process was fairly typical for big tech: HR screening call, 1 hour tech screening (20 minutes background chat and 40 minutes coding challenge), and then either 1 more ML-specific interview for interns or 3 interviews - a mix of coding/ML/system design - for full-time employee candidates.

Q: What do you recommend to fellow tech interviewers?

A: Understand that different people spent a different amount of time on Leetcode or HackerRank, and don't assume that's related to how much the candidate cares about getting the job.

Q: How did you prepare before your first interview as an interviewer?

A: First, I had to read interview guidelines, and complete an online training. Then I became a shadow interviewer - sitting silently in the room to see how more experienced interviewers lead the interviews, and study a bunch of problems. The first couple of times I had a "reverse shadow" interviewer (a more experienced person who stayed silent) to ensure I was ready.

Fun Fact

Hi, I am Kat Stam, the ex-CTO who helps you interview the right tech talent faster.

✳️ I prepare the structured technical interview (questions with answers and live challenges with solutions) for you and teach you how to evaluate each candidate.

✳️ I conduct structured tech assessments across all tech positions, including product, delivering the evaluations and ranked list of candidates straight to your inbox.

None of these sounds useful?

Try self-paced learning here

1. Follow the 2h Tech Interviewer’s Blueprint https://katstam.gumroad.com/l/the-2h-tech-interview-blueprint

2. Learn how to think like a programmer (for non-technies) https://katstam.gumroad.com/l/think-like-a-programmer