Instacart Data Scientist interview experience

I walked through scenarios that mixed some casual machine learning questions with a lot of experimentation-style problem solving, and I had to think hard about how metrics and decisions connected to modeling rather than treating it as purely theoretical.

SQL showed up as well, but the deeper theme was framing: I was expected to organize the problem properly in the relevant domain and stay grounded in what would matter experimentally. The conversations sometimes felt uneven because the interviewer didn’t always seem prepared, which made it harder to read the intended difficulty. Still,