DoorDash·Machine Learning Engineer·Onsite - System Design / Architecture
Jun 2026
DoorDash MLE system design round focused entirely on ads infrastructure. No ML theory, just backend: how you store, retrieve, rank, and pace ads at scale. Harder than I expected to keep the scope tight.
- Design an ads recommendation and serving system, focusing on the backend infrastructure rather than the ML model itself.
- How would you handle budget pacing for advertisers to prevent overspending or underspending within a campaign window?
- How would you design the click and impression tracking pipeline to ensure exactly-once billing?
- Walk through how you would retrieve candidate ads for a given user and context at low latency and high throughput.
- What are the non-functional requirements you would prioritize for an ads serving system, and how do they create tensions with each other?
“This is a beast of a question because it sounds like an ML design problem but they explicitly want you off the model and onto the plumbing.” The rest of the author's notes on Machine Learning Engineer interview at DoorDash, Onsite - System Design / Architecture round, covers how they worked through the question, what the panel pushed back on, and what they would do differently.
View Post