Search the archive.

Chunking decides what your retriever can find. Here's how I split real documents — structure-first boundaries, parent-child units, and how to prove it worked.

Bad RAG retrieval is four or five distinct failures wearing one costume. Here's how I localize which one you have before changing anything.

Exponential backoff caps how fast a client retries; jitter stops every client retrying together. Here's the retry logic I actually ship, and the two limits most teams miss.

A cross-encoder reranker re-scores your top candidates by reading query and document together. Here's how I add one, size it, and prove it worked.

A DLQ is the giving-up mechanism, and most teams build it wrong. The routing, envelope, isolation, and redrive patterns I use to make failed messages recoverable.

Build hybrid search in Postgres with pgvector, tsvector, and RRF in one SQL query — the schema, index tuning, and when you actually need real BM25.