Track your progress 📊

Tick off each block as you finish it. A day is done when all four of its blocks, read, drill, build and write, are ticked.

Your progress is saved in this browser, on this device, and nowhere else. It is not sent anywhere and it is not an account, so it will not follow you to another browser or another machine. Clearing it is one button, below.

Loading your progress...

Week 1: ground truth

Day 1: how slow is slow? R D B W

Day 2: estimation vs reality R D B W

Day 3: what happens when you type a URL R D B W

Day 4: latency, throughput and the queue R D B W

Day 5: designing an API R D B W

Day 6: more than one server R D B W

Day 7: your first real design R D B W

Week 2: storage

Day 8: how a row sits on disk R D B W

Day 9: B-trees and LSM trees R D B W

Day 10: indexes in depth R D B W

Day 11: transactions and isolation R D B W

Day 12: replication and lag R D B W

Day 13: partitioning and sharding R D B W

Day 14: designing a storage layer R D B W

Week 3: caching and the CDN

Day 15: caching and the cache-aside pattern R D B W

Day 16: eviction and the working set R D B W

Day 17: the thundering herd R D B W

Day 18: hot keys and the celebrity problem R D B W

Day 19: Redis internals and pipelining R D B W

Day 20: the CDN and cache invalidation R D B W

Day 21: designing a news feed R D B W

Week 4: async, queues and the log

Day 22: sync vs async, and why queues exist R D B W

Day 23: the log as a primitive R D B W

Day 24: Kafka, partitions and consumer groups R D B W

Day 25: delivery semantics and idempotency R D B W

Day 26: the outbox pattern R D B W

Day 27: backpressure and load shedding R D B W

Day 28: designing a notification system R D B W

Week 5: distributed systems

Day 29: failure, time, and the eight fallacies R D B W

Day 30: CAP, for real, and PACELC R D B W

Day 31: replication and quorums R D B W

Day 32: consensus and Raft R D B W

Day 33: logical clocks R D B W

Day 34: the papers R D B W

Day 35: designing a distributed key-value store R D B W

Week 6: AI systems

Day 36: how LLM inference actually works R D B W

Day 37: the KV cache and continuous batching R D B W

Day 38: vector search and embeddings R D B W

Day 39: RAG, and why it mostly fails at retrieval R D B W

Day 40: agents, and why the token bill explodes R D B W

Day 41: serving concerns, caching, limiting and fallback R D B W

Day 42: designing a production AI chat system R D B W

Week 7: production

Day 43: observability R D B W

Day 44: SLOs and error budgets R D B W

Day 45: rate limiting R D B W

Day 46: multi-tenancy and isolation R D B W

Day 47: security boundaries R D B W

Day 48: cost and capacity planning R D B W

Day 49: designing a multi-tenant SaaS R D B W

Week 8: putting it all together

Day 50: the framework R D B W

Day 51: mock, a URL shortener R D B W

Day 52: mock, a chat system R D B W

Day 53: mock, a rate limiter and a news feed R D B W

Day 54: mock, an AI system R D B W

Day 55: capstone kickoff R D B W

Day 56: capstone build R D B W

Day 57: capstone load test R D B W

Day 58: capstone writeup R D B W

Day 59: review and gaps R D B W

Day 60: the finish R D B W

New days appear here automatically as they are published. The full roadmap for all sixty days lives in the week pages in the sidebar.