← DoorDash Interview Insights

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

SeniorPrefer not to say
May 2026Remote

Summary

DoorDash ML engineer interview focused heavily on a multi-objective ranking problem for restaurants. The whole thing was basically one long design question with a lot of follow-ups baked in, which I wasn't fully expecting.

Questions Asked (5)

Q1

Design a restaurant ranking model that jointly optimizes for conversion rate and expected delivery time. Walk through your candidate architectures.

System DesignTechnical Trade-offsAlgorithms & Data Structures
Author's notes

I went straight to multi-task learning with shared layers and two output heads, which felt like the obvious move.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying the business objectives and constraints, then propose a multi-objective ranking model that combines conversion rate and delivery time into a single utility function. Discuss candidate architectures like multi-task learning and learning-to-rank, and explain how you would train and evaluate them.

Pro tip: Emphasize the importance of online experimentation (A/B testing) to validate offline metrics and tune the trade-off, and mention how you would handle position bias in conversion data.

1. Clarify Objectives and Constraints

Ask questions to understand business goals, such as the relative importance of conversion vs. delivery time, and any constraints like fairness or latency. Define clear metrics for both objectives.

2. Formulate the Problem

Frame it as a multi-objective optimization problem. Propose combining the two objectives into a single utility score, e.g., U = conversion_prob - λ * delivery_time, where λ controls the trade-off.

3. Propose Model Architectures

Describe candidate architectures: (1) Multi-task learning with shared layers and separate heads for conversion and delivery time, (2) Learning-to-rank with a combined loss, (3) Two-stage: first filter by delivery time, then rank by conversion.

4. Discuss Training and Evaluation

Explain how to train the model using historical data, handling biases like position bias. Define offline metrics (e.g., NDCG for ranking, calibration for conversion) and plan for online A/B testing.

5. Address Trade-offs and Scalability

Discuss trade-offs between model complexity and latency, and how to scale to millions of restaurants. Mention techniques like feature engineering, model distillation, or caching.

Key Points to Mention

  • Multi-task learning to share representations between conversion and delivery time prediction.
  • Utility function with a tunable parameter λ to balance conversion and delivery time.
  • Handling position bias in conversion data using techniques like inverse propensity scoring.
  • Offline evaluation metrics: NDCG, AUC for conversion, MAE for delivery time.
  • Online A/B testing to validate and tune the trade-off in production.
  • Scalability considerations: low-latency inference, distributed training, and feature serving.

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

Q2

How would you design the loss function for this two-objective model, and how do you weight the two heads against each other?

Technical Trade-offsProduct Analytics & Metrics
Author's notes

Talked about a weighted sum of the two losses and mentioned tuning the weight as a hyperparameter.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying the two objectives and their business context, then propose a combined loss function (e.g., weighted sum or uncertainty-based weighting) and justify the weighting strategy with data-driven reasoning. Emphasize that the design should align with product goals and be validated through offline metrics and online A/B tests.

Pro tip: Mention that you would treat the loss weights as hyperparameters to be tuned via Bayesian optimization or grid search, and that you'd monitor per-task metrics to avoid negative transfer. This shows you understand both the theory and the practical MLOps workflow.

1. Clarify objectives and constraints

Ask clarifying questions to understand what the two objectives are (e.g., click-through rate and delivery time prediction) and how they relate to business metrics. Identify if there are any hard constraints or priorities.

2. Choose a loss combination strategy

Propose a combined loss function, such as a weighted sum of individual losses, or more advanced methods like uncertainty weighting (Kendall et al.) or gradient normalization. Explain the trade-offs of each approach.

3. Determine weighting scheme

Discuss how to set the weights: start with equal weights or based on business importance, then tune them using validation performance. Mention techniques like grid search, random search, or Bayesian optimization.

4. Validate and iterate

Describe how you would evaluate the model: track per-task metrics, overall business metrics, and check for negative transfer. Use offline evaluation and online A/B testing to refine weights.

5. Monitor and adapt in production

Explain how you would monitor the model post-deployment, detect drift in task importance, and potentially adjust weights dynamically if needed.

Key Points to Mention

  • Weighted sum of losses with tunable weights
  • Uncertainty-based weighting (homoscedastic uncertainty)
  • Gradient normalization or dynamic weight adjustment
  • Business metric alignment and trade-off analysis
  • Per-task evaluation metrics and negative transfer detection
  • Hyperparameter tuning and A/B testing for weight selection

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

Q3

Once you have predictions for both CVR and delivery time, how do you combine them into a single ranking score?

System DesignTechnical Trade-offsProduct Strategy
Author's notes

This is where it got interesting.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying the business objective—whether the goal is to maximize orders, revenue, or customer satisfaction—since that determines the trade-off between conversion and delivery time. Then propose a combined score, such as a weighted sum or multiplicative model, and explain how to tune the weights using online experiments or offline simulations.

Pro tip: Mention that delivery time should be treated as a penalty that grows non-linearly (e.g., exponential) because customers become disproportionately frustrated with long waits. Also, highlight the importance of calibrating predictions and handling missing values to avoid biased rankings.

1. Clarify business objective and constraints

Ask what the primary goal is (e.g., maximize orders, revenue, or retention) and any constraints (e.g., fairness, delivery capacity). This sets the context for how to weight CVR and delivery time.

2. Choose a combination function

Propose a function like score = CVR * f(delivery_time) or score = CVR - λ * delivery_time, where f is a decreasing function. Discuss pros and cons of additive vs. multiplicative models.

3. Determine the trade-off parameter

Explain how to set λ or the shape of f using business rules, historical data, or A/B tests. Emphasize that this parameter controls the balance between conversion and speed.

4. Validate and iterate

Describe offline evaluation (e.g., replay logs, counterfactual estimation) and online experiments (A/B tests) to measure impact on key metrics. Mention the need to monitor for feedback loops and biases.

5. Consider personalization and context

Note that the optimal trade-off may vary by user segment, time of day, or location. Suggest building a model that predicts the trade-off parameter or uses features to adjust the score dynamically.

Key Points to Mention

  • Business objective alignment: conversion vs. delivery time trade-off depends on company goals.
  • Non-linear penalty for delivery time: customers are more sensitive to longer waits.
  • Calibration of CVR and delivery time predictions to ensure comparability.
  • Handling missing or uncertain predictions (e.g., fallback strategies).
  • Online experimentation (A/B testing) to tune the combination weights.
  • Personalization: different users may have different sensitivities to delivery time.

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

Q4

How do you handle label correlation and selection bias when training this model?

Data ModelingRoot Cause Analysis
Author's notes

Blanked for a second here.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying the specific model and data context (e.g., delivery time prediction, demand forecasting) to ground your answer. Then, systematically address label correlation and selection bias by explaining detection methods, mitigation techniques, and validation strategies, emphasizing practical trade-offs. Conclude with how you would monitor and iterate in production.

Pro tip: Tie your answer to business impact: explain how unaddressed biases could lead to poor decisions (e.g., underestimating delivery times in certain regions) and how your approach improves fairness and accuracy. Mention that you'd collaborate with domain experts to validate assumptions.

1. Clarify the problem and data

Ask clarifying questions about the model's purpose, data sources, and how labels are generated. Identify potential sources of correlation and selection bias (e.g., only including completed orders, feedback loops).

2. Detect label correlation and selection bias

Use statistical tests (e.g., correlation matrices, VIF) and visualizations to identify correlated labels. For selection bias, compare distributions of selected vs. non-selected data and check for missing not at random (MNAR) patterns.

3. Apply mitigation techniques

For label correlation, consider multi-task learning, label smoothing, or decorrelation penalties. For selection bias, use reweighting, propensity score matching, or Heckman correction. Choose based on model constraints and interpretability.

4. Validate and monitor

Evaluate model performance on holdout sets that reflect the target population, using fairness metrics and bias audits. Set up production monitoring for drift in label distributions and selection rates.

5. Iterate with stakeholders

Share findings with domain experts to refine assumptions and adjust the approach. Document trade-offs and ensure alignment with business goals.

Key Points to Mention

  • Label correlation: causes (e.g., hierarchical labels, shared features) and impact on model performance (e.g., inflated importance, poor generalization).
  • Selection bias: types (e.g., survivorship bias, sampling bias) and how it can skew training data and predictions.
  • Detection methods: correlation analysis, missing data analysis, and comparing distributions.
  • Mitigation techniques: multi-task learning, reweighting, propensity scores, and data augmentation.
  • Validation strategies: using unbiased holdout sets, fairness metrics, and A/B testing.
  • Production monitoring: tracking label drift, selection rates, and model performance across segments.

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

Q5

What offline metrics would you track for this system, and how would you set up an online A/B test to evaluate it?

A/B Testing & ExperimentationProduct Analytics & Metrics
Author's notes

Pretty standard but I made it harder than it needed to be.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying the system's objective (e.g., delivery time prediction, search ranking) and define offline metrics that align with business goals, such as RMSE for regression or NDCG for ranking. Then outline a rigorous A/B testing plan that includes hypothesis, randomization unit, primary/secondary metrics, sample size calculation, and guardrail metrics to ensure a holistic evaluation.

Pro tip: Emphasize that offline metrics are proxies and may not capture real-world dynamics; propose a pre-A/B validation step like counterfactual logging or a small-scale online pilot to de-risk the full experiment.

1. Clarify system and objectives

Ask clarifying questions to understand the system's purpose, business goals, and constraints. This ensures metrics are relevant and aligned with stakeholder expectations.

2. Define offline metrics

Select offline metrics that measure model performance and correlate with business outcomes. Include both primary (e.g., RMSE, AUC) and secondary metrics (e.g., calibration, fairness).

3. Design online A/B test

Specify the experiment design: hypothesis, randomization unit (e.g., user, order), control/treatment, primary and guardrail metrics, sample size, duration, and statistical tests.

4. Address potential pitfalls

Discuss challenges like network effects, novelty effects, and metric sensitivity. Propose mitigation strategies such as stratification, CUPED, or switchback tests.

5. Conclude with iteration plan

Explain how results will inform next steps, including model retraining, feature iteration, or further experiments. Highlight the importance of continuous monitoring.

Key Points to Mention

  • Offline metrics: RMSE, MAE, AUC, precision/recall, NDCG, calibration, business-specific metrics (e.g., delivery time accuracy).
  • Online metrics: conversion rate, order completion time, customer satisfaction (CSAT), delivery partner utilization, guardrail metrics like app crashes or latency.
  • A/B test setup: randomization unit (user ID, order ID), sample size calculation (power analysis), test duration (accounting for weekly seasonality), and statistical significance (p-value, confidence intervals).
  • Advanced techniques: CUPED for variance reduction, sequential testing, multi-armed bandits for adaptive allocation, and switchback experiments for marketplace settings.
  • Potential biases: novelty effect, primacy effect, network effects, and Simpson's paradox; use holdout groups and long-term metrics to detect.
  • Alignment with business: tie metrics to North Star (e.g., deliveries per hour, customer retention) and ensure offline-online correlation via offline replay or counterfactual evaluation.

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