Vector Database Deep Dive: How They Actually Work
An engineer's guide to embeddings, ANN search, HNSW, IVF, and choosing between pgvector, Pinecone, Qdrant, Weaviate, and Milvus
Search across 221 posts, 23 explainers, and 79 topics
An engineer's guide to embeddings, ANN search, HNSW, IVF, and choosing between pgvector, Pinecone, Qdrant, Weaviate, and Milvus
Inside Spokes, the three-replica system that keeps half a billion repositories online for a hundred million developers
From the moment SELECT leaves your app to the row that lands back, what really happens inside Postgres
The 1978 logical clock that taught distributed systems how to order events without a shared wall clock
The HLC pattern that combines wall clock time with a logical counter to give you causal ordering and human readable timestamps
The log cleanup pattern that keeps your WAL from eating all your disk space
The commit index pattern that prevents your cluster from serving stale or inconsistent data
Architecture, performance, trade-offs, and the decision framework that actually works in production
How to publish messages reliably without distributed transactions, with real implementations using polling and CDC
Architecture, throughput, cost, delivery guarantees, and the real trade-offs developers hit in production
Instrumentation, Collector pipelines, sampling, Kubernetes deployment, and the mistakes that cost teams weeks
How MCP works, why it matters, and how to build your first MCP server
The complete journey from pressing Enter to pixels on screen, explained for software developers
Shared locks, exclusive locks, deadlocks, MVCC, and how to stop your transactions from fighting each other