Production RAG: Chunking, Hybrid Search, Reranking and Evaluation
Design a production RAG pipeline from ingestion to evaluation, with practical decisions for chunking, hybrid retrieval, reranking, permissions and grounded answers.
Read article07 / Journal
Thoughts from buildingreal AI systems.
Notes on LLM and RAG systems, MLOps, data engineering and the craft of shipping software that has to keep working.
- 39
- Articles published
- 21
- Topics covered
02 / All articles
Everything I’vewritten down.
Filter by topic or search for a specific problem.
2 articles of 38
LLMOps in Practice: Evaluate, Trace and Monitor LLM Applications
Connect offline evaluations to production traces and useful alerts. Build an LLMOps process that explains regressions, protects private data and supports controlled releases.
Read articleDiscuss an idea
Working on somethingyou want to talk through?
If one of these posts is close to a problem you have, tell me about it - I’m happy to compare notes.