Advice to the company: Actually design an ML interview if you're trying to hire top ML talent.

The process started with a 30-minute presentation about a past ML-related project, in front of a panel. Almost everyone was late, and only one person introduced themselves. Nobody asked any questions at the end, so a manager (who joined halfway through) had to essentially put somebody on the spot to ask a random question.

This was followed by more general coding questions and a behavioral interview.

During the entire process, nobody asked me a single question related to Machine Learning/AI. Instead, it seems that (except for the hiring manager interview) I was simply thrown into a standard Software Engineering loop, with people from non-ML teams doing the coding/design interviews.

The onsite included a "design" question that was not about designing ML applications (which is what the job is about), but constructing a vaguely defined "scheduler system". This question seemed to have just come out of their general question bank, and there was no understanding that this isn't a question that's super relevant to an ML engineer's work.

I got a thumbs up from all interviews except this one, which is not surprising since I'm not a Software Engineer.

Overall, I'm pretty shocked that a company that's branding itself as wanting to become a big player in AI does not have an interview loop that even acknowledges that building ML/AI applications requires a different skill set.

How would you design a system to:

• schedule jobs
• retry with exponential backoff

No context or use case was given. Minimal help from the interviewer.