5 years of experience interviewing for an ML infrastructure role. Unfortunately, I wasn't able to pass the interview but the interview process was fairly standard.
The call was fairly standard, with the caveat of asking about AI safety. Make sure you develop a strong opinion on AI safety before going into the interview. I believe having a view that is simultaneously pessimistic and optimistic is ideal as the whole company revolves around AI safety as a mission statement - while doing AI
The question is this one (your standard web crawler).
https://leetcode.com/problems/web-crawler/description/
Fairly standard with a BFS solution, but the follow up was multithreading which required the use of thread pool executors and asyncio. I answered all the questions properly, passed all test cases - but I think the interviewer was somewhat biased towards me.
The result was a rejection shortly after in a few days even though I was sure I could've moved to the onsite.