Glossary
Retrieval-Augmented Generation (RAG)
Grounding an LLM's answer in retrieved documents at query time.
What it is
RAG retrieves relevant chunks from your data and adds them to the prompt so the model answers from real sources instead of memory. It reduces hallucination, enables citations, and keeps answers current without retraining.
Why it matters in production
RAG is the most-shipped pattern in applied AI — and the most under-engineered. Demo RAG is an afternoon; production RAG is chunking strategy, hybrid search, reranking, evals, and cost control. That gap is exactly what interviews probe.
In practice
A 'chat with your docs' app: embed document chunks into a vector store, retrieve the top chunks per question, and have the LLM answer only from them with citations.
Go deeper
Related terms
Production AI Notes
One practical AI engineering email each week
One concept, one architecture, one project idea, and one interview question — written for developers who want to build and ship real AI systems.