Interview timeline:
6/15/2026 – DoorDash HR reached out on LinkedIn asking about my availability for an interview.
6/23/2026 – DoorDash first VO round, CodeCraft interview. The question was Dasher Pay — there are already write-ups about this problem, both here and elsewhere. There was a follow-up asking how to handle the double pay rate for dashers during peak hours. After I finished the code, the interviewer asked verbally: what if the upstream API fails? You can answer this with backoff with jitter, circuit breaker, etc. I passed this round smoothly.
7/7–7/8/2026: Virtual Onsite round, four interviews total — AI Code Craft, Debugging Round, Engineering Values + HM talk, System Design + Domain Knowledge.
Overall impression of the interviews: first, each interviewer came from a different team, brought together just for the interview loop. With such different backgrounds among the interviewers, it was hard for me to find common technical ground with them, and they also had a hard time following my past projects.
Day 1, I had SD + HM:
SD: the interviewer had me design a system for rating food. The page shows the average rating and reviews for a given food item. A user can upvote/downvote someone else's review, and can also post their own review — if someone's review gets a lot of upvotes, the system needs to reward them. You need to pay attention to the schema design: how to distinguish between a user posting a review and a user upvoting/downvoting a review, and how to make sure a person can only upvote once / reward once.
HM talk: the interviewer asked a lot about how performance was evaluated at my past jobs, and who sets OKRs and how. They also dug deep into my projects, plus the classic BQ questions: biggest failure, what would you improve if you redid the project, that kind of thing.
Day 2, I had Code + Debug:
Debug: the interviewer gave Part 1, a broken round robin problem, with the background of assigning orders to backend workers. Part 2 was writing consistent hashing — I ran out of time on this one and just wrote pseudocode for add_node, get_node, etc. with the interviewer. The whole time, the interviewer kept pushing me to hurry up and write faster.
AI code craft: the interviewer had me solve a DoorDash order-timeout refund problem — given a DAG, I needed to write a complete workflow and workflow engine. The interviewer gives you a code skeleton; you copy the HackerRank code out into a local IDE and use AI to run it. You can check in with the interviewer on each change before letting the AI run the code. Follow-up: I was asked multiple times about the impact of repeated API calls, and how to shorten user-facing latency.
Overall, in three of the four rounds the interviewers all seemed rushed, constantly pushing me to move faster. The Debug round had basically no real interaction — just the interviewer repeatedly hurrying me along, which was not a great experience. I wasn't in great shape for the Debug round, and the SD design probably wasn't deep enough either. Three days after the interview, I got a rejection.
Discussion
Loading comments…