Initial screening covered availability, location, and team fit. The first round focused on recent publications, with discussion aimed at understanding the candidate’s research background, interests, and alignment with the role.

Recruiter call then online interviews, leetcode style. The virtual interview focused on data structures and deep learning basics. There were 3 interviews in total, whole process took 1.5 months and interviewers were nice.

The first two rounds focused on ML coding and in general resume and ML discussions. The onsite focused on RL coding, running an evaluation of a vision pipeline and a system design round.