Interview practice

ML System Design

End-to-end ML system design questions covering architecture, trade-offs, and production considerations

20 questions · Try the related quiz first

Design an Ad Click Prediction System

seniorLong form
Answer

Design a Fraud Detection System

midLong form
Answer

Design an ML Monitoring System

staffLong form
Answer

Design a Real-Time Recommendation System

staffLong form
Pro

Design a Search Ranking System

seniorLong form
Pro

Investigate a Cancellation Predictor with Suspiciously High Accuracy

juniorLong form
Pro

Restore a Failed Daily Prediction Batch

juniorLong form
Pro

Route Support Tickets with a Human Escalation Path

midLong form
Pro

Evaluate a Shipment ETA with Incomplete Outcomes

midLong form
Pro

Diagnose Training-Serving Skew in a Demand Forecast

midLong form
Pro

Decide Whether a Small Forecast Win Is Ready

midLong form
Pro

Backfill Corrected Features Without Rewriting History

seniorLong form
Pro

Roll Out a Predictive Maintenance Model Safely

seniorLong form
Pro

Serve Bursty Listing Predictions Under a Cost Cap

seniorLong form
Pro

Diagnose Interference in a Marketplace Experiment

seniorLong form
Pro

Train a Sensitive-Data Model with a Deletion Path

seniorLong form
Pro

Respond to a Seasonal Demand Regime Change

seniorLong form
Pro

Set Contracts for Shared ML Features

staffLong form
Pro

Migrate a Fleet of Models to a New Inference Runtime

staffLong form
Pro

Choose Global or Regional Damage-Risk Models

staffLong form
Pro