DoorDash · Phone screen → Onsite
Rejected2026 DoorDash Experienced SWE Phone Screens and Onsites: 2 Experiences
Both candidates opened with Dasher Pay on the phone. Their onsites had the same kinds of rounds but different system design and debugging questions.
Coverage: Two candidates changing jobs, in loops for full-time general software engineering roles at DoorDash, both listing a Master's degree and a headhunter as the channel. Each had a phone screen and four onsite rounds. Both reports date the interviews to October–December 2026 and went up on November 3, 2026, about five hours apart. Neither candidate passed.
What came up
Both loops had the same five kinds of rounds. Only Dasher Pay repeated as a named question.
| Topic | Reports | Most recent report |
|---|---|---|
| Dasher Pay in the phone screen | 2 of 2 | Nov 3, 2026 (2026 Q4) |
| System design | 2 of 2 | Nov 3, 2026 (2026 Q4) |
| Debugging an existing program | 2 of 2 | Nov 3, 2026 (2026 Q4) |
| AI-assisted coding round | 2 of 2 | Nov 3, 2026 (2026 Q4) |
| Hiring manager round | 2 of 2 | Nov 3, 2026 (2026 Q4) |
Questions we found
Algorithms / Data Structures
| Question | Frequency | All last seen dates |
|---|---|---|
| Dasher Pay with Store Waits and Peak Hours | 2 reports | Reports: 2026-11-03 (both, interviews 2026 Q4); additional FastPrep history: 2026-10-31, 2026-10-16, 2026-10-15, 2026-10-02, 2026-09-29, 2026-09-24, 2026-09-18, 2026-09-16, 2026-09-01, 2026-08-03, 2026-07-24, 2024-10-15 |
System Design
| Question | Frequency | All last seen dates |
|---|---|---|
| A three-day, $100M donation drive with a third-party payment provider (Design a Three-Day Charity Donation Drive) | 1 report | Report: 2026-11-03 (interview 2026 Q4); additional FastPrep history: 2026-07-30 |
| Food review + reward (Food Review System) | 1 report | Report: 2026-11-03 (interview 2026 Q4); additional FastPrep history: 2026-10-16, 2026-09-29, 2026-09-24, 2026-09-21, 2026-09-16, 2026-09-01, 2026-08-15, 2026-04-01 |
Debugging
| Question | Frequency | All last seen dates |
|---|---|---|
| A load balancer with buggy round-robin routing, then consistent hashing (Fix a Traffic Router and Add Consistent Hashing) | 1 report | Report: 2026-11-03 (interview 2026 Q4); additional FastPrep history: 2026-10-31, 2026-09-16, 2026-08-15 |
| "Get random" (Random Dasher Set Operations) | 1 report | Report: 2026-11-03 (interview 2026 Q4); additional FastPrep history: 2026-10-31, 2026-09-16, 2026-09-01, 2026-08-03, 2026-04-01 |
AI Coding
| Question | Frequency | All last seen dates |
|---|---|---|
| A refund workflow engine that walks a DAG of nodes (Build a Refund DAG with Local HTTP Services) | 1 report | Report: 2026-11-03 (interview 2026 Q4); additional FastPrep history: 2026-10-31, 2026-09-21, 2026-09-16, 2026-08-15, 2026-08-03 |
Behavioral (BQ)
| Question | Frequency | All last seen dates |
|---|---|---|
| Taking ownership of something beyond your responsibility | 1 report | Report: 2026-11-03 (interview 2026 Q4) |
| Your biggest failure and what you'd change | 1 report | Report: 2026-11-03 (interview 2026 Q4) |
| Improving something that already worked ("1% better") | 1 report | Report: 2026-11-03 (interview 2026 Q4) |
| Making room at the table or helping a teammate grow | 1 report | Report: 2026-11-03 (interview 2026 Q4) |
| A technical disagreement and how it was resolved | 1 report | Report: 2026-11-03 (interview 2026 Q4) |
| Hands-on use of AI tools: where they fell short, and where AI could help teams most | 1 report | Report: 2026-11-03 (interview 2026 Q4) |
| Redoing a past project from a product angle: what would you change, and why? | 1 report | Report: 2026-11-03 (interview 2026 Q4) |
| If you joined a new team, what would you ask team members about AI? | 1 report | Report: 2026-11-03 (interview 2026 Q4) |
| If you had an E3 and an intern, how would you divide the work? | 1 report | Report: 2026-11-03 (interview 2026 Q4) |
2 detailed candidate timelines
01: A vague Dasher Pay, a refund engine and a two-day onsite
Interview period: 2026 Q4 · Report date: November 3, 2026 · Role: Software engineer (general), full-time · Degree listed: Master's · Candidate's rating: Neutral, average difficulty · Result: Did not pass
Round 1: Code Craft, phone screen, 1 hour
- Dasher Pay, given as a day's stream of order events. Each event carries an order ID, a dasher ID, a time and a status, and the listed statuses are
ACCEPTED,FULFILLEDandCANCELED. Pay builds up by the minute for each order active at the same time, and you return the day's total. FastPrep's practice version is Dasher Pay with Store Waits and Peak Hours. Its event names and store-wait rule are its own, and it has no cancel event. - The candidate calls the prompt vague on purpose. Before coding, you have to settle how overlapping orders get paid, whether a cancelled order earns anything, how events are sorted and what the input looks like.
- The main follow-up adds a double-pay window, with 5–7pm as the example, so any interval crossing the window's edge has to be cut in two.
- After the code, the talk moved to production: making the batch script a live service, and handling upstream API failures. The candidate lists backoff with jitter, circuit breakers and idempotency.
The onsite was four rounds across two days.
Round 2 (day 1): AI Code Craft, 1 hour
- The format: your own laptop and AI editor, with the screen shared. HackerRank handled handing the code in and out.
- The task: build the engine behind an automatic refund workflow for late or timed-out orders. Starter code already defines the workflow as a DAG. Its nodes include checking the delay, issuing the refund and notifying, and each node's outcome picks the next edge. FastPrep's practice version is Build a Refund DAG with Local HTTP Services.
- Follow-ups: partial refunds, extra node types such as cancel, the effect of repeated API calls, and cutting the wait users see when the workflow is slow.
- To close, the interviewer wanted a recap of the session (problem, design choices, tradeoffs, where the code stood) and an ASCII architecture diagram labeling components, data flow and protocols.
Round 3 (day 1): System design, 75 minutes
- Roughly the first 20 minutes went to a project the candidate had led, with hard follow-ups on its architecture, the alternatives they'd weighed and problems they'd solved.
- Then the design: a three-day, $100M donation drive where a third-party provider takes the payments. It had to cover the donation flow, the provider integration, instant acknowledgement, and getting scale and money correctness right inside a fixed window. FastPrep's practice version is Design a Three-Day Charity Donation Drive.
Round 4 (day 2): Debugging, 1 hour
- A load balancer exercise in Python on HackerRank:
Pod,Backend,HttpRequestandTrafficRouterclasses, a unittest suite, and round-robin code with bugs. The job is to fix it so traffic spreads across pods that are up and skips those that aren't. You can practice this as Fix a Traffic Router and Add Consistent Hashing. - Part 2 swaps the routing for a consistent-hash ring: one user keeps landing on one pod, and a pod joining or leaving moves little traffic. The stated aim is a better hit rate for the in-memory session cache.
Round 5 (day 2): Hiring manager and engineering values, 45 minutes
Behavioral in STAR style, built around DoorDash's values. Seven of the questions in the BQ table above come from this round, and the candidate singles out the one about redoing a past project as a recurring DoorDash question. The round also touched on their background, how performance and OKRs worked at their last job, and some general chat.
02: A fast-interrupting design round and a muted debugging interviewer
Interview period: 2026 Q4 · Report date: November 3, 2026 · Role: Software engineer (general), full-time · Degree listed: Master's · Candidate's rating: Negative, average difficulty · Result: Rejected
The candidate says HR found them in a wide net on LinkedIn, so they gave the loop a try.
Phone screen
The classic Dasher Pay, with no further details. You can practice it as Dasher Pay with Store Waits and Peak Hours.
Onsite, four rounds
System design. A known question, food review + reward, which FastPrep has as Food Review System. The candidate's complaint is about pacing. They'd planned to build up from simple to deep, but this interviewer jumped in with follow-ups straight away, and every interruption made them look unprepared and scrambling. Partway through, they switched to leading with their best answer. They came away thinking the round went okay, and they note that other candidates have also called DoorDash interviewers very pushy.
Debugging. A known question, the "get random" one, which FastPrep has as Random Dasher Set Operations. The practice version isn't a debugging exercise: you write the set from an empty stub, the input is a fixed String[] of operations, and PICK draw uses a given number instead of a random pick, so there's no code to repair. The interviewer stayed on mute doing other work, gave perfunctory answers and, the candidate felt, didn't know the question well. The costly moment came when the candidate asked whether changing the input type would break the tests and was told it should be fine. Tens of minutes went by before they found it couldn't be changed, and one test case still failed at the end.
AI Codecraft. Another known question, though the report doesn't describe the task. While the AI was implementing, the talk turned to why they chose this tool and this model and whether they'd written skills or agents at work. The last question was what they'd optimize if the system went to production. They finished what was asked and ran an end-to-end test, but they still aren't sure what the round was meant to test.
Hiring manager. The deeper questions included what they'd ask team members about AI after joining a new team, and how they'd split work between an E3 and an intern. Most of the round was basic BQ. The hiring manager was friendly and talkative, and the candidate says this was the only round that felt good.
Afterward
A rejection came a few days later. The candidate asked for feedback and got no reply.
Patterns across the reports
-
Dasher Pay opened both loops. Reports 01 and 02 both had it as the phone screen. 01 gives the details, including the peak-hour follow-up. 02 only calls it the classic Dasher Pay.
-
The onsites shared round types, but the questions they name differ. Both had a system design round, a debugging round, an AI coding round and a hiring manager round. System design was a donation drive in 01 and food review + reward in 02. Debugging was a load balancer in 01 and "get random" in 02. 01's AI round was the refund engine, and 02 doesn't describe its task.
-
AI came up with the hiring manager too. 01 was asked about hands-on use of AI tools and where they fell short. 02 was asked what they'd ask team members about AI after joining a new team.
Comments
0No comments yet — be the first to share what you know.