I recently completed the onsite with Meta in London for the E5 ML position. I'm still waiting for the feedback. In the meantime I want to share my experience. I won't be able to share specific details because of the NDA.

Phone screen: This was my first interview with a FAANG company, so I was very nervous. The interviewer was super friendly thought, and they turned the interview into a pleasant conversation. 2 medium questions (both tagged)

Onsite:

Coding 1: 2 medium questions (1 tagged, 1 not tagged). I quickly found optimal solutions for both problems and (I think) I explained the logic well. I made a few stupid mistakes while implementing one of the solutions. I corrected a few of them before the interviewer noticed, but he pointed one problem which I then quickly corrected. For the second problem the implementation was correct, but the complexity I provided wasn't quite right.

Coding 2: 2 medium questions (both tagged). I quickly found optimal solutions for both problems. I couldn't implement the optimal solution for one of the problems because of the time constraint, but the interviewer seemed happy with the non-optimal implementation.

ML System design: Build ML system for one of the FB products. I have ML background so this was the easiest one from my perspective and I think I did well.

Backend System design: I haven't done any backend system design interviews before, so I was worried about this one. I was trying to follow a systematic structure which I practiced. But I think nerves got into me and I didn't provide a clear description of the data models and made a questionable decision for one of the databases. So, I have very mixed feelings about this one.

Behavioural: Standard questions. I had responses for all questions, but thinking more about it I'm worried that I wasn't giving E5 signal in some of my answers.

During the process the recruiter was very responsive and they provided me with a lot of prep material. I'll post an update once I know the outcome. Any thoughts on what the outcome may be?