Jump to a problem, path or page
Learning paths
Each path orders problems so earlier ones build the vocabulary later ones rely on. Your active path drives the dashboard's next suggestion.
Core request paths, durable state, caches and async side effects on approachable problems. Finish this before the advanced HLD paths.
Entities, state machines, replaceable policies and contracts you can test. The classic machine-coding round problems.
Hot keys, invalidation, edge delivery and precomputation for systems where reads dominate writes.
Move work off the critical path safely: delivery guarantees, ordering, backpressure, dead letters and replay.
Overselling, double charging and split-brain: atomic claims, idempotency keys, fencing tokens and sagas.
High-volume ingestion, time-series and log storage, search indexes, analytical queries and batch movement of data.
Keep systems available while regions fail, shards move, schemas change and configuration rolls out to a fleet.
Caches, heaps, tries, filters and balanced trees, each with explicit invariants and complexity targets.
Single-process components that stay correct under concurrent callers, timeouts, cancellation and failing dependencies.
Streaming decoders, grammars, routers and graph algorithms with clear error contracts and bounded resources.
Jump to a problem, path or page