I made 2 interviews: screening with a recruiter and a live coding one, with Leetcode like exercises without any AI coding assistant. In the coding one they required a very deep knowledge of Python libraries, some of them not common at all in the day to day of a data scientist

Awful. Red flags for a toxic work culture. I met 3 interviewers, all were exhausted. God knows what they were building. Completely drained and engrossed into their own world, asking theoritical to most difficult questions, kind of to reject you. And even if you pass, there is not a good vibe. They pay a penny and expect you to a coder, Ml researcher and Deep learning Gen AI architect working at OpenAI or Google Deepmind ! Shame on your process. No idea it was coming.

1. Screening w/ HR, salary expectations, general questions 2. Algos (2 leetcode medium, one DP) 3. ML theory w/ old stuff, SVM, cross entropy derivation etc. Some math "coding" in excalidraw 4. More advanced ML, stopped here 5. ? final discussion / culture fit ?