Uber Data Scientist Interview Experience — Two Tech Screens, Screwed Over by My Own Round 2 Interviewer

Two rounds of tech screens at Uber — tl;dr, I got screwed over by one of my own.

Sharing this for those who come after.

Round 1: coding and case study. The coding part was simple Python scripts, nothing like an LC medium, I blew through it in 15 minutes. The case study asked about the various downstream effects of a change in rider ETA, and how I'd design an experiment for it — we went from supply-side and demand-side success metrics, to intermediate metrics, to synthetic control, to a classic A/B test design. Had a good back-and-forth with the interviewer the whole way, felt great.

Round 2 was where things went off the rails. Something felt off right from the self-introduction — "I don't understand how it works." The case study was about the impact of a rider incentive, a similar kind of framework — I brought up rider / driver / matching quality and tried to engage the interviewer on which angle they wanted me to dig into, but they were checked out the whole time: "I don't know, you tell me." "Ok, what else, what other metrics." No hints at all about what they actually wanted. Were they stuck on some analysis project and fishing for ideas? They'd keep asking the same question again right after I'd already answered it, then scribble notes furiously.

I don't want to get into too much more detail, but basically it came across like they thought I was a DS with no structure, no logic — the kind of vibe where they'd say "I don't care how the feature is designed, just tell me how you'd measure it."

The job market is hard enough already. I wasn't even asking for a leg up from someone who's one of my own — just didn't expect to get screwed by one instead. Fine, I guess.