DoorDash Software Engineer (E4) Interview Experience
DoorDash · Mid level · Software Engineer
The screen gave me only an empty main function. I had to write everything, including reading the input, and that tripped up people I know more than the logic did.
Interview process
The structure was pretty standard: recruiter screen, then a technical screen, then the onsite loop. The surprising part for me was that once I got past the screen, there were no pure DSA rounds in the onsite at all. The most unusual round was the debugging one, because I started by fixing a round robin implementation with test cases and then had to switch gears into consistent hashing. So even though the loop looked normal from the outside, the actual rounds felt a bit different from the usual big tech pattern. That was the main thing that stood out to me.
Interview rounds · 5
- 1
Recruiter screen
BehavioralProject DiscussionArtificial IntelligenceIt was a pretty standard recruiter screen. He mostly checked my background, the kind of work I had done, whether my stack matched, and a couple of logistics things.
Q1. Can you walk me through your previous experience, the tech stack you worked in, and your domain knowledge?
How they answeredI walked through the kind of work I had done before, the tech stack I had worked in, and the domain knowledge I had. It felt like a normal alignment call to see whether my background matched what they needed.
Q2. How do you use AI in your day-to-day work?
How they answeredI explained how I use AI in day-to-day engineering work as a practical productivity tool. It was not super deep, but they clearly wanted to know whether I already use it in a real workflow.
Q3. Are you okay with relocation?
How they answeredI answered the standard logistics questions around relocation and similar basics. Nothing tricky there.
- 2
Technical round
CodingTechnicalThe screen did not feel like a normal fill-in-the-function DSA round. They gave me only the main function and I had to build the whole thing from scratch, including input handling, which can trip people up if you only practice on platforms with boilerplate.
Q1. Given dasher activity timestamps, calculate the payout for the day based on the delivery-time formula.
How they answeredI had to write the whole program myself, not just complete a helper function. The problem was about a delivery partner's activities, like when an order was accepted and when it was fulfilled, and there was a formula for how much payout should be given for the time spent delivering across the day. So I handled the input parsing and the core logic end to end. I think that part catches people because even syntax and basic I/O matter here.
Follow-up questions- Implement everything from scratch, including how the program takes input.
- 3
Final / onsite round
DebuggingCodingTechnicalThis round was on HackerRank and felt very practical. I was given working code plus a few test cases, had to run them, find the bugs, and then they added a more open-ended follow-up.
Q1. Run the provided test cases on this round robin implementation and find the bugs.
How they answeredI ran the three or four test cases against a small round robin code snippet and fixed a few minor bugs. The issues were things like indexing not being done correctly and assignment to the list not being done correctly. It was pretty straightforward once I executed the cases and traced the failures.
Follow-up questions- What specifically is broken in the code?
Q2. Now convert that round robin logic to consistent hashing.
How they answeredAfter the debugging part, they asked me to convert that piece from round robin to consistent hashing. There were no hard constraints, so I was free to structure it however I wanted, and I wrote more of a pseudocode-style implementation while explaining the approach.
- 4
Final / onsite round
System DesignExecutionThe system design round was backend-focused and pretty grounded. Instead of staying generic, they pushed into payment correctness and what happens when a short-lived event creates a traffic spike.
Q1. Design the backend system for a donation app that will only be live for a short while.
How they answeredI designed the backend for a donation app that would be live for about three days. I started with the functional and non-functional requirements, then walked through the request path and how the payment flow would work. In the deep dive, they focused on how I would prevent double payments and how I would handle peak traffic during that short burst. So the important part was showing payment correctness and scalability.
Follow-up questions- What are the functional and non-functional requirements?
- How will the request flow and payment flow work?
- How would you restrict double payments?
- How would you handle peak traffic?
- 5
Final / onsite round
BehavioralProject DiscussionArtificial IntelligenceThe hiring manager round was mostly behavioral. He wanted to understand how I work through team changes, how I think about feedback, and whether I could go deep on a project from my resume.
Q1. How did your team handle it when a senior teammate left?
How they answeredI talked through how our team handled losing a senior teammate, what I did to keep things moving, what feedback I had received, and how I think about giving feedback upward. He seemed to care about how I operate when the team dynamic changes, not just whether I can tell a polished story.
Follow-up questions- What feedback did you receive?
- What feedback do you give to more senior people?
Q2. Tell me about this project on your resume and what you specifically did.
How they answeredHe picked one project from my resume that he was very interested in, and I went into the domain, what the project was about, and what I specifically worked on. He was trying to get past the headline and see whether I had real depth there.
Q3. How do you use AI in your day-to-day work?
How they answeredI answered again how I use AI in my day-to-day workflow. Since it came up here and in the recruiter round, it felt like they wanted a practical answer, not just a buzzword answer.
Tips from the candidate
I would definitely practice coding in something like IntelliJ with no autocomplete, because here they can make you write the full thing from scratch, even the input handling, and some people get slowed down by that more than by the actual logic. For system design, do not stop at the high-level architecture. Be ready to talk through double payments, idempotency, and what happens when traffic spikes hard for a very short event. And for the debugging round, practice reading someone else's code, running tests, fixing concrete bugs, and then still being ready for a follow-up that becomes more open-ended.
Company culture
They seem to like realistic backend scenarios more than abstract LeetCode-style puzzles. A lot of the process lined up with high-traffic and scalability-heavy work, which makes sense for DoorDash, and that is actually one reason the role felt interesting to me. AI also came up more than once, so it feels like they care whether engineers are already using it in a practical day-to-day way. From what I know, the people are helpful, but there is healthy peer pressure too.