Skip to content

Learning paths

All links are existing pages. Status on each title is whatever project status says today — do not treat a Needs review page as gold.

How to move through any path: How to Study. The method itself: Design framework.

The method (do not skip)

Four canonical modules are the intended loop — requirements → numbers → bottleneck → smallest justified architecture → failure → evolution. If you only have a week, do these four properly instead of skimming twenty.

  1. URL shortener — V1 is one primary; cache/CDN are earned.
  2. Rate limiter — cardinality and fail-open vs fail-closed before Redis folklore.
  3. Circuit breakers — timeout → retry math → breaker; run the storm sim.
  4. Sharding — hot key vs more boxes; then consistent hashing.

Short voice-over of the same loop: Reasoning transcripts. Score designs with assessment.

Senior Backend

Goal: derive a service, store, cache, and async path without cargo-culting.

  1. Framework + estimation + math
  2. Stateless vs stateful · API design (Needs review — use for HTTP contracts, not as Complete)
  3. SQL vs NoSQL · Indexing · Sharding
  4. Cache strategies · Cache stampede
  5. Queue patterns · Kafka
  6. Rate limiter · URL shortener · Notification system
  7. Circuit breakers · Debugging playbook
  8. LLD: Parking Lot + SOLID (principles page is Needs review; the exercise is Complete)

Staff Distributed Systems

Goal: replication, consensus, partitions, and evolution under failure.

  1. Canonical four above
  2. CAP · Consistency models · Replication · Raft
  3. Distributed fundamentals · Multi-region DR
  4. WhatsApp · Payments · Distributed KV
  5. Sagas · CRDTs · Event-driven architecture
  6. Failure library · Failure-injection tasks
  7. Social feed or Collaborative editor

Platform / Infrastructure

Goal: traffic path, clusters, change, cost.

  1. HTTP & TCP · Load balancing
  2. Docker · Kubernetes · K8s kind lab
  3. Deployment strategies · FinOps
  4. CI/CD · Terraform (both Needs review — labs still useful: Terraform+Docker)
  5. Load balancer exercise · API gateway
  6. IAM & managed services · Testing strategy
  7. Code deployment / release orchestration

Data / Streaming

Goal: logs, consumers, batch vs stream, hot partitions.

  1. Queue patterns · Kafka · Kafka lab
  2. Canonical: Sharding + Kafka hot-key injection task
  3. Stream processing · Event sourcing & CQRS
  4. DDIA concepts · Indexing
  5. Distributed message queue · Log aggregation · Metrics & monitoring
  6. Batch/ETL (Needs review) — prefer stream-processing + DDIA until it is rewritten
  7. Kafka internals / Pulsar (Needs review) after Kafka consumer groups is solid

Interview Sprint

Goal: 10–14 days to be fluent, not encyclopedic.

Days 1–2. How to Study · Framework · Estimation · Math

Days 3–5. Canonical four (shortener, limiter, breaker, sharding) + CAP + Raft

Days 6–8. WhatsApp · Payments · Kafka · Cache stampede

Days 9–10. DSA: Sliding window · BFS/DFS · DP · Pattern recognition

Days 11–12. Technical disagreement · Production incident · Debugging playbook

Days 13–14. One mock scored with assessment; failure-injection on Kafka + Raft + breaker; skim cheat sheets

Production Debugging

Goal: symptom → hop → mechanism.

  1. Debugging playbook (p99 + Kafka lag)
  2. Kubernetes request path + K8s sim
  3. Tail latency · Circuit breakers · Retry-storm lab
  4. Failure library · Failure-injection tasks
  5. Replication + Postgres lab
  6. Cache stampede · Single points of failure (Needs review — still the right prompt: “point at the one box”)
  7. Production incident · Production reliability practices

After a path

Mark pages on Your Progress. If a module is Needs review, you may still read it — just do not confuse notes with a finished learning loop. Track honesty on the quality matrix.