Online coding assignment. Very theoretical, the few practical questions are very easy. Most of them don't really assess much about your ML knowledge. If that is passed, then you're invited for a "technical" interview. They call it ML triage, there's recommended ML books on their website (they haven't a clue about those) but questions were not ML related at all, but standard leetcodes instead. The interviewer also had no background in ML. They then send you a vaguely belittling feedback from your interviewer (with no ML knowledge) about your ML interview (with no ML in it) about why you (an ML PhD with multiple big tech internships) are not a good fit for an ML research internship with them. To say it is a waste of time is reductive, since the content of the interviews are 100% dependent on who's interviewing you.

I can’t discuss specifics of questions asked as I signed an NDA. First three technical interviews went OK and the interviewers seemed genuinely interested in talking to me and presented the questions in an engaging way. They gladly discussed some interesting projects that I was interested in working on. Technical questions were relevant to the projects. Final stage interview was supposedly a “culture fit interview”. Instead, I was interviewed by senior management who asked technical questions irrelevant to the project I had been assigned to at G Research. Interviewers were terse and did not seem to be interested in tracking my progress of the question, rather just ask the question, wait a few minutes and state the answer if I haven't got there yet. One of the interviewers was on his phone for much of the interview, completely ignoring me. Feedback for the final interviews was sent to me and seemed to be written in a disingenuous and deliberately belittling manner, as though they were trying to make fun of me. Unsurprisingly I didn't receive an offer - no problem for me as I have other offers & I would prefer not to work with people who enjoy putting down others.