Retrieval-Augmented Generation (RAG) — From Fundamentals to Production-Ready Agentic RAG Systems

advanced40+ hours54 sections

A comprehensive, end-to-end course through Retrieval-Augmented Generation — beginning with core concepts and document processing, advancing through embeddings, vector stores, and retrieval techniques, and culminating in agentic RAG systems built with LangGraph and a deployable capstone project.

What you'll learn

Understand the RAG architecture and when to use it vs fine-tuning vs prompt engineering
Master document processing, chunking strategies, and metadata management
Select and implement appropriate embedding models for different use cases
Deploy and operate vector stores including Chroma, FAISS, Qdrant, and Pinecone
Implement basic and advanced retrieval techniques including hybrid search, MMR, and re-ranking
Build advanced RAG patterns: RAG Fusion, HyDE, Corrective RAG, Self-RAG, and Graph RAG
Design and implement agentic RAG systems using LangGraph with multi-step reasoning
Evaluate RAG pipelines using RAGAS with comprehensive metrics
Deploy a production-ready RAG system with monitoring, caching, and optimization

Prerequisites

Python fundamentals (functions, classes, async/await)
GenAI fundamentals (LLM basics, tokens, prompting)
LangChain basics (Chains, LCEL, tool usage)
LangGraph basics (StateGraph, nodes, edges, conditional routing)

Course outline

1. RAG Fundamentals and Architecture

2. Document Processing and Chunking

3. Embeddings and Vector Representations

4. Vector Stores

5. Basic Retrieval Techniques

6. Advanced Retrieval Techniques

7. Advanced RAG Patterns

8. Agentic RAG with LangGraph

9. Evaluating RAG with RAGAS

10. Capstone Project with Deployment