System Design track
AI & ML System Design Interview Questions
Curated interview questions. Solve them in a guided workspace, then get your design scored against a rubric.
ML and LLM system design, plus quick AI concept revision.
AI concepts: quick revision25 short questions with model answers: embeddings, RAG, KV cache, quantization, prompt injection, evals and more.0/25 revisedBeginner0/3
Intermediate0/6
Advanced0/7
—
sign in to track
16 questions
| Status | Question | Asked at | Best | Time |
|---|---|---|---|---|
| Design a Spam Classifier Beginnerclassification · feature-engineering · class-imbalance | GoogleMicrosoftMeta+1 | — | 45m | |
| Design a Sentiment Analysis Service Beginnernlp · fine-tuning · batch-vs-online | AmazonFlipkartMicrosoft+1 | — | 45m | |
| Design Product Search Ranking (Basic) Beginnerlearning-to-rank · retrieval · click-data | AmazonFlipkartMeesho+2 | — | 45m | |
| Design a Recommendation System (YouTube / Netflix) Intermediatetwo-tower · candidate-generation · ranking | GoogleNetflixAmazon+2 | — | 60m | |
| Design News Feed Ranking Intermediatemulti-task-learning · ranking · integrity | MetaLinkedInTwitter+1 | — | 60m | |
| Design a RAG-based Document Q&A System Intermediaterag · chunking · vector-db | OpenAIAnthropicMicrosoft+3 | — | 60m | |
| Design Semantic Search with a Vector Database Intermediateembeddings · ann · hnsw | GoogleAmazonPinecone+2 | — | 60m | |
| Design a Fraud Detection System Intermediateanomaly-detection · real-time-features · class-imbalance | PhonePeRazorpayPaytm+3 | — | 60m | |
| Design Ad Click-Through Rate Prediction Intermediatectr-prediction · calibration · sparse-features | GoogleMetaAmazon+2 | — | 60m | |
| Design a ChatGPT-like LLM Chat Service Advancedllm-serving · streaming · rate-limiting | OpenAIAnthropicGoogle+2 | — | 75m | |
| Design an LLM Inference Platform Advancedgpu-scheduling · continuous-batching · kv-cache | OpenAIAnthropicGoogle+2 | — | 75m | |
| Design an AI Coding Assistant Advancedcode-completion · context-retrieval · latency | GitHubMicrosoftAnthropic+2 | — | 75m | |
| Design a Multi-agent / Agent Platform with Tool Calling Advancedagents · tool-calling · orchestration | OpenAIAnthropicMicrosoft+2 | — | 75m | |
| Design an LLM Evaluation & Guardrails Pipeline Advancedevals · llm-as-judge · guardrails | OpenAIAnthropicGoogle+2 | — | 75m | |
| Design Real-time Personalization Advancedstreaming-features · session-based · bandits | AmazonNetflixSpotify+2 | — | 75m | |
| Design an Image Generation Service Advanceddiffusion · gpu-queues · content-safety | OpenAIGoogleAdobe+2 | — | 75m |
Sign up to save drafts, submit designs and track your scores.