The interview process consists of 8 interviews: 2 phone, and 6 onsite. In a phone interview, one was generic ML, and another one was more ML depth plus little coding(ML related). On onsite interviews there were 6 rounds, most of them were ML depth, ML coding focused except one leetcode focused. Though everyone on the panel was exceptionally good, the interview process itself was a bit faulty. Interviews were scattered all over places, generally Coding, ML depth happens in the first half, but in my process, it was all haphazard. My last round was leetcode focused and after 5 hours of rigorous interviews it was tough to focus and it didn't go that well.

I progressed through multiple interview stages for a research role and was invited to attend onsite interviews. Based on the confirmed schedule, I arranged my travel plans in advance, including planning train tickets and organizing my stay to ensure I could attend the in-person round without delays. However, shortly before the scheduled interview, I was informed that the role had been put on hold and the interview was cancelled.Earlier communication regarding the possibility of the role being paused, or confirming role stability before inviting candidates onsite, would significantly improve the candidate experience. On a positive note, the technical interview conversations prior to this were engaging and professional, and I appreciate the time invested by the interviewers. I hope the process can become more predictable and transparent for future candidates.