Applied Systems Lab
A six-week path from failure drills to working mental models — labs included.
Applied Systems Lab
Build judgment for systems that fail in public: caches, leases, feeds, uploads, counters, schedulers, and spikey product paths.
Each lesson follows the same arc — business constraint → naive v0 → failure drill → evolution → implementation cut → numbers → self-check. Pattern cards are reusable tools; labs are small Go packages under labs/ in this repo.
Progress checkboxes save only in your browser (localStorage). No accounts.
Week 1
Week 1: Foundations under load
Presence, pools, cache storms, async pipelines, and log streams — control load before it controls you.
Week 2
Week 2: Coordination and realtime
Locks, leases, store shapes, messaging hubs, and WebSocket scale.
Week 3
Week 3: Edge, media, and counters
Load balancers, leaders, CDN media, upload pipelines, private media, hot counters.
Week 4
Week 4: Storage primitives and agents
Bitcask dictionaries, object stores, RAG grounding, and research agent budgets.
Week 5
Week 5: Orders, events, and LSM
Multi-tier orders, ingestion, LSM flushes, distributed IDs, hot shards.
Week 6
Week 6: Live product spikes
Commentary feeds, schedulers, flash sales, impressions, and matching.