First I got a reach-out from a recruiter, and we set up a quick call to go over the basics and the interview process. Then they scheduled the first technical round. After that round, the process got handed off to another recruiter, who also set up a quick call to go over some basic info, and at the same time scheduled the onsite.
The first round was a Code Craft round, 1 hour.
The second round was the onsite — four back-to-back interviews in total (when scheduling you could pick to do them all in one day, or split across two days): AI Code Craft, Debugging, System Design, and HM + BQ.
Code Craft: They gave some template code with three services — user service, payment service, address service — and you had to design an aggregator service that calls all three. I felt like the most important thing was getting the code running and leaving enough time for the interviewer to ask follow-ups, then having one or two questions ready to ask back at the end — that should be enough to pass.
AI Code Craft: They gave a problem description, and you had to open your local IDE and an AI command line — basically however you use AI to write code at your actual job, that's how you're supposed to use it in the interview. The problem was a DAG workflow, and you had to implement it in code so it could automatically issue refunds to users. The follow-up was how to handle partial refunds — also done with AI, definitely doable. But you probably need to talk through with the interviewer what the code you generated is actually doing. There was another follow-up: if this workflow runs slowly and users have to wait a long time, how do you fix that? You could also use AI for this — AI gave 10 different solutions, and looking at them on the spot made me a little dizzy. I went through them one by one explaining which ones were more doable given the existing code, and I don't think I explained that part very well — probably lost some points there (maybe I should have just asked the AI which one was more doable?).
Debugging: I had some bad luck in this round — the interviewer came in almost 10 minutes late, and I couldn't be bothered to reschedule, which probably made the whole debugging session feel rushed. The problem was debugging round robin and consistent hashing. The round robin debugging took a while — there was a string missing a letter that I couldn't find for a long time, which wasted some time. Then for consistent hashing I barely managed to write out a code skeleton and didn't have time to actually run it before the next interviewer walked in, so this round ended pretty abruptly.
System Design: This round was pretty standard. We started with small talk and introductions, then they asked about a project I'd worked on before, with a lot of follow-ups — you really need to prepare for this part. There was one specific example about how I solved a hard problem in the project that I didn't answer well, which was on me for not being prepared enough, and I felt like it wasn't going great at that point. Then the actual system design question was a food review system, where users can rate and review food, and other users can upvote/downvote reviews. I felt like I hadn't done enough mock interviews for design, so there were things I thought of afterward that I could have said but forgot to mention. This round didn't go that well.
HM + BQ: This round was fairly standard. It really did feel like the interviewer was just reading questions off the screen one after another — there was a lot less interaction than the other rounds. But we ended up having a pretty pleasant conversation by the end, and it wrapped up on a relaxed note.
I got the rejection one or two days after the onsite. I've barely seen anyone else on the forum get an offer either, and some people are saying there's a hiring freeze — I don't know why the recruiter was pushing the process so hard. Does DoorDash have almost no headcount but is still actively running interviews anyway?
Discussion
Loading comments…