← DoorDash Interview Insights

DoorDash·Machine Learning Engineer·Onsite - System Design / Architecture·Senior

Senior
Jun 2026

Summary

DoorDash ML Engineer onsite focused entirely on designing a homepage recommendation system. One long, dense question that basically covered everything from retrieval to ranking to cold-start. Felt like a mini system design marathon.

Questions Asked (1)

Q1

Design a personalized homepage recommendation system that retrieves and ranks a feed for each user, covering requirements, data sources, retrieval and ranking stages, feature infrastructure, serving, training, evaluation, and strategies for cold-start and diversity.

System DesignTechnical Trade-offsA/B Testing & Experimentation
Author's notes

This thing sprawled fast.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying requirements and constraints, then walk through the end-to-end ML system: data sources, retrieval, ranking, feature infrastructure, serving, training, evaluation, and cold-start/diversity strategies. Emphasize trade-offs and how you would measure success with offline metrics and online A/B tests.

Pro tip: Anchor your design to DoorDash's business goals—e.g., optimizing for order completion and user retention—and explicitly discuss how you'd handle the cold-start problem for new users and restaurants, since this is critical for a marketplace.

1. Clarify Requirements and Constraints

Ask about scale (users, items), latency, personalization goals, and business metrics. Define functional and non-functional requirements.

2. Data Sources and Feature Infrastructure

Identify user, item, and interaction data. Describe feature engineering, storage (e.g., feature store), and real-time vs. batch features.

3. Retrieval and Ranking Stages

Explain candidate generation (e.g., collaborative filtering, embeddings) and ranking (e.g., learning-to-rank with GBDT/NN). Discuss multi-stage funnel.

4. Serving, Training, and Evaluation

Cover model serving (low-latency, scalable), training pipeline (retraining frequency), and evaluation (offline metrics like NDCG, online A/B tests).

5. Cold-Start and Diversity Strategies

Address new user/item cold-start (e.g., content-based, contextual bandits) and diversity (e.g., re-ranking, MMR) to avoid filter bubbles.

Key Points to Mention

  • Two-stage architecture: retrieval (candidate generation) and ranking (personalized scoring).
  • Feature store for consistent online/offline features and low-latency serving.
  • Cold-start solutions: use side information, explore-exploit (bandits), and fallback to popularity.
  • Diversity techniques: intra-list similarity, category coverage, and business rules.
  • Evaluation: offline (AUC, NDCG) and online (CTR, conversion, A/B testing).
  • Scalability and latency considerations: caching, approximate nearest neighbors, model quantization.

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.