You can find my phone screen writeup in a separate post.
Since this was a recruiter reach-out, once I passed the phone screen I went straight to onsite with the team that wanted to interview me.
The onsite was split into the following parts:
- Collaboration interview — we talked about the current SOTA in my research direction, its pros and cons, whether I had relevant technical experience, and how I saw the field developing going forward, plus how to apply these research directions to product.
- System design — design an end-to-end real-time sensor system. "End-to-end" meant starting from sensor selection, the use case, and the PM's request, then having to consider algorithm choice, how to design the ML model, where the data comes from, and efficiency and power. This system design leaned heavily on domain knowledge and was tied to the specific team I was interviewing with.
- Research presentation — present your own recent research results. I presented a paper and got asked a lot of motivation, why, and how questions. They also challenged whether my experimental design was sound.
- Team lunch — just chatting about what everyone was working on.
- Team chat #1 — similar to a behavioral round: how do you collaborate with others, what do you do when you hit difficulties, how do you work cross-functionally.
- Team chat #2 — also behavioral-ish, same as above, but I also got asked about some technical experience.
- HM interview — mainly about future projects, and whether the team's direction was a good match.
Other than the research presentation I didn't prepare much — I skimmed some papers in the field. Everyone was very nice and the overall experience was good. I think the onsite topics were pretty random and hard to prepare for, but I'd read some recent papers in the field; since I've been doing product work for a while and haven't done much research, I needed to spend time catching up.
Got the offer about a week later.
Discussion
Loading comments…