Tracks

Follow a track, not a feed

Each track is a curated sequence — a learning path through one corner of the systems map. Pick one and go deep.

Consensus, replication, ordering, failure — the hard parts, made intuitive.

Indexes, transactions, isolation — what's actually happening under the query.

From the TCP handshake to the TLS one, and everything a request touches.

Races, locks, backpressure, memory models — the bugs that only show up under load.

Auth, crypto, threat modelling — the attacks, the defences, and the tradeoffs.

Embeddings, inference, retrieval, fine-tuning — building systems that actually work.

HTTP, browsers, CDNs, caching — everything between the user and your server.

Latency, throughput, profiling, SLOs — making systems fast and keeping them up.

Metrics, traces, logs — and how to actually find the thing at 2am.

CI/CD, containers, orchestration, IaC — shipping fast without breaking prod.