Course Wiki

RAG

Retrieval-augmented generation: ground an LLM's answers in retrieved documents instead of relying on what it memorized.

RAG is the architecture LLM Zoomcamp is built around. A user question is answered in three steps: retrieve relevant documents from a knowledge base (search over your own data), assemble them into a prompt with the question, and have the LLM generate an answer grounded in that context. The model explains your documents instead of guessing from its training data.

Module 1 builds the pipeline in its simplest form — keyword search over a FAQ dataset, a prompt template, an OpenAI API call — then spends the course hardening each stage: better retrieval with vectors, answer quality measurement in the evaluation module, and feedback loops in monitoring.

Taught in

• LLM Zoomcamp

  • Module 1: Agentic RAG
    • Introduction
    • RAG
    • The Course FAQ Dataset
    • Search
    • The LLM
    • RAG Helper
    • Data Ingestion
    • Wrap-up of Part 1
    • Agents
    • Quick RAG Revision (Optional)
    • Function Calling
    • Other Frameworks
  • Module 2: Vector Search
    • Vector Search
    • Vector Search with minsearch
    • RAG with Vector Search
    • Vector Search with sqlitesearch
    • Vector Search with PGVector
    • Next Steps
  • Module 3: Orchestration
    • AI Orchestration
    • Retrieval Augmented Generation
    • Best Practices
    • Next Steps
  • Module 4: Evaluation
    • Evaluation
    • Search Evaluation
    • Search Parameter Tuning
    • RAG and Agent Evaluation
    • Generating RAG Answers
    • LLM as a Judge
    • Agent Evaluation
    • Next Steps
  • Module 5: Monitoring
    • Monitoring
    • Assistant
    • Chat App
    • Capturing Metrics
    • Storing Data in PostgreSQL
    • User Feedback
    • Built-in Judge
    • Docker Compose
    • Next Steps
  • Module 6: Best Practices
    • Best Practices for RAG
    • Next Steps
  • Module 7: End-to-End Project
    • End-to-End Project Example
    • Evaluating Retrieval
    • Evaluating RAG
    • Interface and Ingestion Pipeline
    • Monitoring and Containerization
    • Summary and Closing Remarks
    • Chunking for Longer Texts
    • Content Processing Cases and Steps ## Related concepts

• Agentic RAG

• Keyword Search

• Embeddings

• Vector Search

• LLM Evaluation