Roblox Data Scientist Interview Experience — Nailed the Case Study, Still Rejected After Round One

(QS 50-100 school, physics PhD, desperately job hunting.)

First round, technical interview, 60 minutes. The interviewer asked about my resume for 5 to 10 minutes first (my own work also has some connection to DS).

Then two simple pandas data-processing questions using a bit of basic statistical test knowledge. I didn't need to run test cases, I just had to make the interviewer understand my code. I didn't finish the last step, so the interviewer just had me talk through what I wanted to do and how, which was fairly simple, and I got it all out.

After that it was verbal. They asked a case: how do you determine whether a variable affects how long users spend on the product. I used DID, found the confounder, and designed a method to handle it. I got through this one quickly and the interviewer said "nice" several times.

Then they asked some ML questions, all fairly basic, and I answered those too. When time was almost up, the interviewer asked if I had extra time because they wanted to ask a few more questions (I took this as a positive signal), and of course I said yes. Then they asked me some linear regression questions, and I answered all of those as well.

At the end, the usual "do you have any questions for me."

My own read was that I did fine overall (maybe a bit slow on the coding?) — I answered the case, the ML questions, and the linear regression questions, and the interview ran 70 minutes. (I believe my answers were correct because the interviewer clearly asked a lot of follow-ups.)

But I didn't even pass round one. I don't know why I got rejected. It's demoralizing — I don't know what level you have to be at to pass these interviews.

Asking everyone here for some advice.