1 machine learning phone screening, followed by all day interviews with a 1 hour presentation and six one-on-one interviews. Interview is easy in the sense you don't get asked super technical questions. I had a single non-algorithmic coding problem. The only tricky part is that whenever you talk about your machine learning or data science work, you must be prepared to justify why you did what you did. If you don't give a good justification, the interviewer uses this as evidence you don't approach problems using the appropriate ways, and hence you aren't qualified for the position. Interview overall was mixed. First interviewer kept referencing her "score" sheet out loud as I would answer each question, e.g. "Oh so you didn't perform extrinsic evaluation (check)," which made it appear she was trying to find evidence against my candidacy from the start and made this interview feel antagonistic. Next interviews were much better, and I felt with each technical question I was working with the interviewers on solving it like with a peer, rather than simply being tested. Two interviewers stood out as people I'd like to work with. Interview with hiring manager was also like the first interviewer, in that it was more of a test than some collaborative process, and hence I think it didn't go well. The recruiter got back to me the next business day, saying while I am a good fit for Amazon's culture/leadership principles and quote "did very well", I had mixed feedback on the technical side.