← DoorDash Interview Insights

DoorDash·Machine Learning Engineer·Technical Phone Screen·Senior

Senior
Apr 2026

Summary

DoorDash ML engineer interview, open-ended case style where they drop a domain problem on you and expect a full walk-through from framing all the way to monitoring. No coding, just a long conversation about how you'd actually build the thing.

Questions Asked (7)

Q1

How would you frame an ML problem in the DoorDash domain (like ETA prediction or dasher dispatch), and what metrics would you use to measure success?

Product Analytics & MetricsProduct Sense & Ideation
Author's notes

This is where I spent too long hedging.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying the business objective and translating it into a precise ML problem, then outline the modeling approach and define success metrics that align with both offline and online evaluation. Emphasize the trade-offs between different metrics and how they tie back to DoorDash's key performance indicators like delivery time and dasher efficiency.

Pro tip: Show that you understand the two-sided marketplace: optimizing for one side (e.g., customers) can hurt the other (e.g., dashers), so propose metrics that balance both and mention guardrail metrics to prevent regressions.

1. Clarify business objective

Ask clarifying questions to understand the specific goal, such as reducing delivery time or improving dasher utilization, and the constraints (e.g., real-time, data availability).

2. Formulate ML problem

Translate the objective into a formal ML task: define the prediction target (e.g., ETA in minutes), input features (e.g., distance, traffic, restaurant prep time), and output type (regression, classification, ranking).

3. Choose modeling approach

Propose suitable models (e.g., gradient boosting for ETA, reinforcement learning for dispatch) and discuss how to handle challenges like real-time inference, scalability, and cold start.

4. Define success metrics

Select offline metrics (e.g., MAE, RMSE for ETA; total delivery time for dispatch) and online metrics (e.g., average delivery time, dasher utilization, customer satisfaction), and explain how they link to business KPIs.

5. Address trade-offs and evaluation

Discuss trade-offs between metrics (e.g., accuracy vs. latency), propose A/B testing or online evaluation, and mention guardrail metrics to monitor unintended consequences.

Key Points to Mention

  • Business objective alignment: ETA prediction improves customer experience and reduces support calls; dasher dispatch optimizes delivery efficiency and dasher earnings.
  • Problem formulation: ETA as regression with features like distance, traffic, restaurant prep time; dispatch as a sequential decision problem or assignment optimization.
  • Offline metrics: MAE/RMSE for ETA, total delivery time or dasher utilization for dispatch; use of holdout sets and cross-validation.
  • Online metrics: average delivery time, on-time delivery rate, dasher utilization, customer ratings, and order completion rate.
  • Trade-offs: balancing accuracy and latency, fairness in dispatch, and exploration vs. exploitation in real-time systems.
  • Guardrail metrics: monitor for regressions in customer satisfaction, dasher earnings, or system latency.

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

Q2

What data sources and features would you use to build a model for this problem?

System DesignTechnical Trade-offs
Author's notes

Felt okay here.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying the problem and the model's objective, then systematically outline the data sources (internal and external) and features (raw, derived, real-time) that would drive predictions. Emphasize how each source and feature ties to business impact and technical feasibility, and discuss trade-offs like latency, cost, and data quality.

Pro tip: Anchor your answer in DoorDash's three-sided marketplace (consumers, merchants, dashers) and show how features from each side interact—this demonstrates systems thinking and domain awareness that interviewers value.

1. Clarify the problem and objective

Ask questions to understand the specific prediction task (e.g., delivery time, demand forecasting) and the business goal (e.g., improve customer experience, optimize logistics). This ensures your data and feature choices are relevant.

2. Identify internal data sources

List relevant data from DoorDash's platform: order history, user interactions, merchant data, dasher data, and operational logs. Consider both batch and real-time sources.

3. Consider external data sources

Mention external data like weather, traffic, holidays, and local events that can impact predictions. Discuss how to integrate them (e.g., APIs, third-party vendors).

4. Engineer features from each source

Propose raw and derived features (e.g., time since last order, merchant prep time, dasher speed, weather severity). Include real-time and historical features, and discuss feature stores for consistency.

5. Discuss trade-offs and validation

Address trade-offs: data freshness vs. cost, feature complexity vs. interpretability, and potential biases. Explain how you'd validate feature importance and monitor drift.

Key Points to Mention

  • Internal data: orders, user profiles, merchant menus and prep times, dasher locations and delivery history
  • External data: weather, traffic, local events, holidays, and demographic data
  • Feature types: temporal (time of day, day of week), geospatial (distance, density), behavioral (order frequency, ratings), and real-time (dasher location, current demand)
  • Data pipeline considerations: batch vs. streaming, latency requirements, and feature store for online/offline consistency
  • Trade-offs: cost of data acquisition, feature engineering effort, model complexity, and potential privacy concerns
  • Evaluation: offline metrics (AUC, RMSE) and online A/B testing, plus monitoring for data drift and feedback loops

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

Q3

Walk through your model selection process: what baselines would you start with, and when would you consider moving to deep learning?

Technical Trade-offsAlgorithms & Data Structures
Author's notes

I went gradient boosting baseline first, which I think was the right call.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by framing model selection as a trade-off between performance, latency, and maintainability, then walk through a progressive approach from simple baselines to complex models. Emphasize that you only move to deep learning when simpler models fail to meet business requirements or when the data scale and complexity justify it. Use a concrete example from your experience to illustrate the decision points.

Pro tip: Mention that you always establish a simple baseline first (e.g., logistic regression or gradient boosting) to set a performance floor and to understand the data, and that deep learning is considered only when you have sufficient data, clear evidence of non-linear patterns, and the infrastructure to support it.

1. Define the problem and success metrics

Clarify the business objective, constraints (latency, interpretability, budget), and evaluation metrics (e.g., AUC, RMSE, business KPI). This guides the choice of models.

2. Start with simple baselines

Begin with heuristic or linear models (e.g., logistic regression, linear regression) and tree-based models (e.g., random forest, gradient boosting) to establish a performance baseline and understand feature importance.

3. Iterate and analyze errors

Evaluate baseline performance, analyze misclassifications or residuals, and check if the errors suggest non-linear relationships or interactions that simpler models cannot capture.

4. Consider deep learning when justified

Move to deep learning if: (a) baselines underperform, (b) you have large amounts of unstructured data (images, text, sequences), (c) the problem requires complex pattern recognition, and (d) you have the infrastructure and expertise to train and deploy.

5. Validate and compare trade-offs

Compare deep learning models against baselines on performance, latency, cost, and maintainability. Choose the simplest model that meets requirements, and consider hybrid approaches.

Key Points to Mention

  • Start with simple, interpretable baselines (e.g., logistic regression, decision trees) to set a performance floor and understand the data.
  • Use tree-based ensembles (e.g., XGBoost, LightGBM) as strong baselines for tabular data, often outperforming deep learning.
  • Move to deep learning when data is unstructured (images, text, audio), when there's a need for representation learning, or when simpler models plateau.
  • Consider data size: deep learning typically requires large datasets to avoid overfitting and to leverage its capacity.
  • Evaluate trade-offs: deep learning may offer higher accuracy but at the cost of latency, interpretability, and infrastructure complexity.
  • Mention the importance of iterative experimentation and error analysis to guide model selection.

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

Q4

How would you handle training and evaluation methodology, particularly around avoiding data leakage or temporal issues?

Technical Trade-offsProduct Analytics & Metrics
Author's notes

Temporal train/test splits came up and I got this right, but I blanked for a second on how to handle the case where labels are delayed (like actual delivery time only known after the fact).

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by acknowledging that training and evaluation must mirror the production environment, especially for time-dependent problems like demand forecasting or delivery time prediction. Then walk through a structured methodology: define the prediction time, split data temporally, engineer features only from past data, and use time-based cross-validation. Finally, emphasize continuous monitoring and retraining to prevent leakage in production.

Pro tip: Mention that even seemingly innocuous features like 'average delivery time in the last hour' can leak future information if not computed with proper time windows. Always validate that your feature pipeline uses only data available at prediction time.

1. Define the prediction time and target

Clarify exactly when the prediction is made and what the target variable is. This sets the boundary for what data is available and prevents using future information.

2. Split data temporally

Use a time-based split (e.g., train on older data, validate on newer) instead of random splits. For evaluation, use rolling or expanding window cross-validation to respect temporal order.

3. Engineer features with time-awareness

Compute features only from data prior to the prediction time. Use window functions or point-in-time joins to avoid leakage from future events.

4. Validate and monitor for leakage

Perform sanity checks: if a feature has unrealistically high importance or model performance is too good, suspect leakage. Monitor feature distributions and model performance in production.

5. Iterate and retrain

Regularly retrain models on new data and re-evaluate with temporal splits. Update feature pipelines to reflect changing data patterns and avoid stale leakage.

Key Points to Mention

  • Temporal validation techniques: time series split, walk-forward validation
  • Point-in-time correctness for feature engineering
  • Avoiding target leakage by excluding future-dependent features
  • Use of pipelines to enforce consistent preprocessing
  • Monitoring for data drift and leakage in production
  • DoorDash-specific examples: delivery time prediction, demand forecasting, dasher assignment

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

Q5

What deployment considerations would you think through, specifically around latency and keeping the model predictions fresh?

System DesignTechnical Trade-offs
Author's notes

Latency I handled fine, talked about precomputing some features and serving from cache.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by framing the problem around DoorDash's real-time delivery marketplace, where both latency and freshness directly impact user experience and operational efficiency. Then walk through a layered architecture—from data ingestion to model serving—and discuss trade-offs between latency, freshness, and cost, proposing concrete techniques like streaming pipelines, online learning, and caching with TTLs.

Pro tip: Emphasize that freshness and latency are often in tension, and show how you'd measure and monitor both (e.g., prediction staleness, p99 latency) to make data-driven decisions. Mention that you'd start with a simple batch approach and iterate based on business impact, demonstrating pragmatism.

1. Clarify Requirements and Constraints

Ask about the specific use case (e.g., ETA prediction, search ranking) and its latency SLA (e.g., <100ms) and freshness needs (e.g., real-time traffic, restaurant prep time). Identify business impact of stale predictions.

2. Design Data Pipeline for Freshness

Propose a streaming architecture (e.g., Kafka, Flink) to ingest real-time events (orders, driver locations) and update features in an online feature store. Discuss trade-offs between push vs. pull and exactly-once vs. at-least-once semantics.

3. Model Serving and Latency Optimization

Choose a serving pattern: real-time inference with low-latency model (e.g., distilled, quantized) or precomputed predictions with caching. Discuss techniques like model quantization, ONNX runtime, and edge caching to meet latency SLAs.

4. Balance Freshness and Latency with Trade-offs

Explain how to decide between batch, micro-batch, and online learning based on how quickly the underlying data distribution changes. Propose hybrid approaches: e.g., batch retraining daily + online fine-tuning for critical features.

5. Monitoring and Iteration

Define metrics for latency (p50, p95, p99) and freshness (prediction age, feature staleness). Set up alerts and A/B tests to validate that improvements in freshness/latency actually move business metrics.

Key Points to Mention

  • Streaming data pipelines (Kafka, Flink) for real-time feature updates
  • Online feature store (e.g., Feast, Tecton) to serve fresh features with low latency
  • Model serving optimizations: quantization, pruning, ONNX, TensorRT, caching
  • Trade-offs between batch, micro-batch, and online learning for model freshness
  • Latency budgets and SLA adherence (p99 latency, tail latency mitigation)
  • Monitoring prediction staleness and drift, with automated retraining triggers

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

Q6

How would you design an A/B test to evaluate your model in production?

A/B Testing & Experimentation
Author's notes

Straightforward for me relative to the rest of the interview.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying the business objective and the specific model change you want to evaluate, then outline a rigorous experimental design that includes randomization, metrics, and statistical analysis. Emphasize the importance of guardrail metrics and practical considerations like sample size and duration to ensure valid, actionable results.

Pro tip: In production ML, it's crucial to monitor not just the primary metric but also model-specific metrics like latency and prediction distribution shifts, as these can reveal issues that aggregate business metrics might miss.

1. Define Objective and Hypotheses

Clearly state the business goal (e.g., increase orders) and formulate a testable hypothesis about how the new model will impact key metrics. Identify the primary metric and any secondary or guardrail metrics.

2. Design the Experiment

Determine randomization unit (e.g., user, session), traffic split, and sample size based on power analysis. Ensure control and treatment groups are comparable and decide on the duration of the test.

3. Implement and Monitor

Deploy the model to the treatment group, ensuring proper logging and monitoring of both business and system metrics. Watch for any anomalies or technical issues that could invalidate the test.

4. Analyze Results

Use statistical tests (e.g., t-test, bootstrap) to compare metrics between groups, checking for significance and practical impact. Segment analysis can reveal heterogeneous effects.

5. Decide and Iterate

Based on results, decide whether to roll out, iterate, or abandon the model. Document learnings and consider follow-up experiments to refine further.

Key Points to Mention

  • Randomization unit and avoiding contamination between groups
  • Sample size calculation and statistical power
  • Primary, secondary, and guardrail metrics (e.g., business KPIs, latency, error rates)
  • Statistical significance vs. practical significance
  • Potential pitfalls: novelty effects, seasonality, network effects
  • Monitoring model performance and data drift during the test

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

Q7

How would you monitor the model after deployment, and what failure modes would you watch for?

Root Cause AnalysisSystem Design
Author's notes

Went through distribution shift, label delay, and upstream feature pipeline failures.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by outlining a layered monitoring strategy covering data, model, and business metrics, then discuss specific failure modes like data drift, concept drift, and system issues. Emphasize proactive detection and mitigation, and tie your answer to DoorDash's real-time logistics context.

Pro tip: Mention the importance of setting up automated alerts with thresholds based on historical variance, and having a rollback plan ready. This shows you think about operational resilience, not just detection.

1. Define Monitoring Layers

Describe monitoring at three levels: input data quality, model predictions, and business outcomes. This ensures comprehensive coverage from raw data to impact.

2. Identify Key Metrics

List specific metrics for each layer, such as feature distribution stats, prediction latency, accuracy, and business KPIs like delivery time or order completion rate.

3. Set Up Alerts and Dashboards

Explain how you would implement real-time dashboards and automated alerts for anomalies, using tools like Prometheus, Grafana, or custom solutions.

4. Enumerate Failure Modes

Discuss common failure modes: data drift, concept drift, upstream data issues, model staleness, and infrastructure problems. Relate them to DoorDash's dynamic environment.

5. Plan for Mitigation

Outline response strategies: retraining triggers, fallback models, and rollback procedures. Highlight the importance of a feedback loop for continuous improvement.

Key Points to Mention

  • Data drift and concept drift detection using statistical tests (e.g., KL divergence, PSI)
  • Model performance monitoring with ground truth delays (e.g., using proxy metrics)
  • Business metric monitoring (e.g., delivery time, order completion rate) and their correlation with model performance
  • Infrastructure monitoring (latency, throughput, error rates) and its impact on model serving
  • Alerting thresholds and anomaly detection techniques
  • Retraining pipelines and automated rollback mechanisms

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