System Design Interview Checklist (With a Practice Plan)

Minute 0–5: clarify before you draw

Restate the product goal, users, and must-have features. Ask about read/write ratio, consistency needs, latency targets, and multi-region requirements. Interviewers grade structure as much as diagrams.

• Functional requirements (core user flows)
• Non-functional requirements (scale, latency, availability)
• Out of scope items (explicitly park them)
• Success metrics for the system

Minute 5–10: rough capacity

Estimate DAU, QPS, storage growth, and payload sizes. Order-of-magnitude math is enough. The goal is to justify caches, partitions, and async work later.

Minute 10–25: high-level design

Draw clients, API gateway, services, data stores, and major queues. Name components and data flow for the primary use case first. Only then branch into secondary flows.

Minute 25–40: deep dives and failure modes

Pick two hotspots: for example cache invalidation, fan-out, search indexing, or leader election. Discuss bottlenecks, single points of failure, retries, idempotency, and monitoring.

Always end with trade-offs: what you would change at 10× traffic and what you would not build on day one.

A four-week practice plan

Week 1: URL shortener, pastebin. Week 2: news feed, chat. Week 3: rate limiter, job queue. Week 4: timed mocks with feedback. Record yourself once — filler words and skipped requirements become obvious.

Pair checklist practice with Learnisim AI conceptual learning and mock interviews so vocabulary and delivery improve together.