Agentic RAG is powerful but it's simpler than you might think. Here's an explanation and demo.

Description

My curriculum of AI courses: https://edwarddonner.com/curriculum I get so many questions about Agentic RAG! Most of my AI Engineering courses use Agentic RAG in one form or another. It's actually a really simple idea. We build a traditional RAG pipeline, then upgrade it to Agentic RAG in just a few lines of code. We start with a single-agent architecture, then split out a Retrieval Agent / Research Agent. The Agent automatically takes care of advanced RAG techniques like reranking, query rewriting and query decomposition as needed, with no extra code. The repo is here - see the README for full setup instructions: https://github.com/ed-donner/agentic-rag