🏛️ System Design Roadmap: core → advanced

Level 1: Fundamentals (Phases 1–2)

  • Client–server, HTTP, REST, stateless services
  • Vertical vs horizontal scaling; load balancers
  • SQL vs NoSQL; indexes; replication basics
  • Caching layers (browser, CDN, app, DB); cache-aside
  • Estimation Cheatsheet

Level 2: Building blocks (Phases 2–3) → Building Blocks

  • Load balancer, API gateway, reverse proxy
  • Cache (Redis), CDN, object storage (S3)
  • Message queue vs event log (SQS/RabbitMQ vs Kafka)
  • Search index, rate limiter, ID generator, distributed lock
  • Blob + metadata split; presigned URLs

Level 3: Distributed systems (Phase 3) → Distributed Systems

  • Replication, partitioning/sharding, consistent hashing
  • Consistency models, CAP/PACELC, quorums
  • Consensus (Raft), leader election, clocks
  • Idempotency, exactly-once illusions, ordering

Level 4: Architecture patterns (Phase 3) → Architecture Patterns

  • Monolith → modular monolith → microservices
  • Event-driven, CQRS, event sourcing, saga, outbox
  • DDD strategic design (bounded contexts, context maps)
  • Hexagonal / clean architecture (inside a service)
  • Resilience: timeouts, retries, circuit breakers, bulkheads, load shedding

Level 5: Production & scale (Phases 4–6)

  • Multi-region, active-active vs active-passive, cell-based architecture
  • Observability, SLOs, capacity planning
  • Zero-downtime deployments, schema migrations (expand/contract), feature flags
  • Cost-aware architecture
  • AI systems architecture (RAG, agents, LLM gateways) → AI Engineering Roadmap

🎤 Interview framework (45–60 min)

StepTimeOutput
1. Requirements5 minFunctional (3–5 core features), non-functional (scale, latency, availability, consistency)
2. Estimates3 minQPS (avg/peak), storage/yr, bandwidth; only estimate what drives decisions
3. API5 minKey endpoints/events
4. Data model5 minEntities, keys, access patterns → DB choice
5. High-level design10 minBoxes and arrows covering every functional requirement
6. Deep dives15 minThe 2–3 hardest parts (contention, hot keys, fan-out, consistency)
7. Wrap-up5 minBottlenecks, failure modes, monitoring, evolution

What interviewers score

Structured approach · clear trade-offs · depth in 1–2 areas · handling of failures · communication. Not the number of boxes.

Resources

  • ⭐ DDIA · System Design Interview Vol 1 & 2 (Xu) · Understanding Distributed Systems (Vitillo)
  • ⭐ Hello Interview (written deep dives + guided practice) · ByteByteGo · system-design-primer (GitHub)
  • YouTube: Jordan Has No Life, Hello Interview, ByteByteGo, Gaurav Sen (intro)
  • Real architectures: Engineering Blogs & Papers
  • Practice: Case Studies (use the System Design Case template)