Topic reading guide
RAG in production
Retrieval is one boundary in a larger system. Read these essays in order to connect ingestion and permissions to the evidence a generated answer can actually support.
Start with the system
RAG in Production Is Mostly Not About Retrieval
See why ingestion, lifecycle, answer behavior and evaluation matter as much as the search call.
Essay
Design the corpus and query path
Designing a Production RAG System for 50 Million Documents
Follow a complete production design across versioned ingestion, hybrid retrieval, permissions, freshness, index generations and operations.
Essay
Context and evidence
Designing the Context Layer for Production AI Agents
Understand how retrieved candidates are authorized, checked for freshness and assembled into evidence the model may use.
Essay
The Truth Is Not in the Docs
See what changes when engineering knowledge spans code, schemas, incidents and ownership rather than documents alone.
Essay
Evaluate system behavior
Your RAG Evals Are Measuring the Wrong Thing
Evaluate evidence, permissions, freshness and abstention across the path, not only whether the final answer sounds right.
Essay
Keep reading
Continue with a guided sequence of free production engineering essays.
Find your next reading path