DoorDash Data Scientist phone screen — I got new questions on both the case study and the SQL sections.
The case study asked what the impact would be if a delivery is late, and what the possible reasons are. Then, based on the reasons I gave, the interviewer asked what data we would look at, and how to decide which of my stated reasons was the most important one. They also brought up that we've noticed food delivery takes longer than package delivery — why might that be, and how would you design an experiment if we wanted to change it?
The SQL section gave two tables, one an orders table and one a request table, then asked three questions, including things like the percentage of orders in the past 15 days where the arrival time exceeded one hour. My feeling is this was harder than the common "frequent customer" question you usually see on the forum — closer to the difficulty of Meta's onsite SQL questions. There's someone on the forum who put together a summary of Meta's SQL questions, and I think practicing with that would leave you in decent shape even if you run into new questions.
Like a lot of people on the forum have said, I think DoorDash's case interviews are genuinely hard.
Discussion
Loading comments…