RAGs to Rich AIs with LLM Agents
A deep dive into retrieval-augmented generation and LLM agents, taught to a class of 27 students through PESU I/O, PES University's peer-teaching program.
Instructor: Vyoman Jain and Samarth P
Term: Fall
Location: PES University, EC Campus
I taught this course as a Subject Matter Expert through PESU I/O, PES University’s peer-teaching program, to a class of 27 students.
The course covers retrieval-augmented generation from the ground up and then pushes past the naive version of it:
- RAG architecture - chunking, embeddings, vector stores, and why the retrieval step is usually where things break
- Advanced RAG techniques - reranking, query rewriting, hybrid search, and evaluation
- GraphRAG - building and querying knowledge graphs when flat similarity search isn’t enough
- LLM agents - building agentic systems with LangChain, LlamaIndex, and CrewAI
The emphasis throughout was hands-on: students built working systems rather than only reading about them.