I applied through other source.

The first part of the interview was conducted by the AI model Zara. It was scoped to the job description and generally not too difficult. The second part is the coding interview where I was given a task to solve. That also is easy to medium LeetCode level. The review I had said I should have explained my thinking process. I absolutely had no idea and thought the final answer was what was of interest since it was an AI interview.

How do you avoid leakage in preprocessing pipelines with sklearn?

How do you ensure determinism in a multi-distributed GPU setting?