New York Site views —Page views —
← All series

Building production LLM systems

Read the sequence on context, retrieval, evidence and abstention.

Reading order

  1. Beyond prompt engineering: building a reliable context system
  2. Context engineering: combining prompts, retrieval, memory and tools
  3. Compressing and ordering context so a model can use the evidence
  4. Evaluating RAG retrieval: diagnosing recall, ranking and attribution
  5. Citations and attribution in RAG: tracing claims back to evidence
  6. RAG ingestion and chunking: turning documents into retrievable evidence
  7. Hybrid retrieval and reranking in RAG: combining keywords, vectors and a reranker
  8. Low confidence in RAG: when to answer, clarify or hand off