All resources
Technical30 min read

System design basics

A step-by-step framework for system design rounds — from requirements to scaling trade-offs.

System design rounds intimidate engineers who have only done feature work, but they reward a calm, structured approach more than encyclopedic knowledge. The interviewer wants to see you scope, design, and reason about trade-offs out loud — not recite a memorised architecture.

A framework that always works

Drive the conversation through these steps instead of jumping to a diagram.

  • Clarify requirements — functional (what it does) and non-functional (scale, latency, consistency).
  • Estimate — back-of-envelope QPS, storage, and bandwidth to justify your choices.
  • Define the API — a few core endpoints frame the whole design.
  • High-level design — client, load balancer, services, database, cache, queue.
  • Deep-dive — the interviewer will pick one component; go deep on data model and bottlenecks.
  • Address scale — sharding, replication, caching, and the trade-offs each forces.

Know the fundamental trade-offs

You do not need to have built a planet-scale system; you need to reason about the levers. SQL versus NoSQL, strong versus eventual consistency, the implications of CAP, when to add a cache and how to invalidate it, and how a message queue decouples producers from consumers. For every choice, say what you gain and what you give up — naming the cost is the signal.

Practise the classics out loud

Work through the canonical prompts — URL shortener, a news feed, a rate limiter, a chat system, a ride-hailing backend — and talk through each end to end as if interviewing. Time-box yourself to 40 minutes. The skill being tested is structured communication under a clock, not the existence of the perfect design.

Key takeaways

  • Drive a fixed framework: requirements → estimates → API → design → deep-dive → scale.
  • Reason about trade-offs (SQL/NoSQL, consistency, caching) — name what each costs.
  • Rehearse the classics out loud in a 40-minute box.

Put it into practice

Run an AI mock interview and get honest, real-time feedback. Free — bring your own AI key.

Start a practice interview