Agentic_RAG

If your RAG retrieves the wrong chunks, your answer can be confidently wrong. And the LLM won't tell you. I built a practical LangGraph implementation that makes the RAG system check its retrieval, correct it and retry when necessary. Free version: https://github.com/ChandulaSenevirathna/Agentic_RAG Advanced version: https://chandula7.gumroad.com/l/Advanced_RAG_LangGraph_Patte... The advanced version includes Corrective RAG, Adaptive RAG, Human-in-the-Loop, Multi-Agent Supervisor, Parallel Fan-Out and Plan-and-Execute patterns.

What happens when your RAG system retrieves the wrong documents? Or when the retrieved context is not enough to answer the question? A traditional RAG pipeline usually doesn't think twice and its path is so its a single attempt generated answer. "Retrieve → Generate → Answer" What I built Retrieve → Reason → Verify → Correct → Answer I created a collection of self contained notebooks demonstrating different Agentic RAG patterns with LangGraph. Each notebook focuses on a practical pattern that you can understand, experiment with and adapt to your own AI projects. Free version: https://github.com/ChandulaSenevirathna/Agentic_RAG Advanced version: https://chandula7.gumroad.com/l/Advanced_RAG_LangGraph_Patte...

What happens when your RAG system retrieves the wrong documents? Or when the retrieved context is not enough to answer the question? A traditional RAG pipeline usually doesn't think twice and its path is so its a single attempt generated answer. "Retrieve → Generate → Answer" What I built Retrieve → Reason → Verify → Correct → Answer I created a collection of self contained notebooks demonstrating different Agentic RAG patterns with LangGraph. Each notebook focuses on a practical pattern that you can understand, experiment with and adapt to your own AI projects. Free version: https://github.com/ChandulaSenevirathna/Agentic_RAG Advanced version: https://chandula7.gumroad.com/l/Advanced_RAG_LangGraph_Patte...

What happens when your RAG system retrieves the wrong documents? Or when the retrieved context is not enough to answer the question? A traditional RAG pipeline usually doesn't think twice and its path is so its a single attempt generated answer. "Retrieve → Generate → Answer" *What I built* Retrieve → Reason → Verify → Correct → Answer I created a collection of self contained notebooks demonstrating different Agentic RAG patterns with LangGraph. Each notebook focuses on a practical pattern that you can understand, experiment with and adapt to your own AI projects. Free version: https://github.com/ChandulaSenevirathna/Agentic_RAG Advanced version: https://chandula7.gumroad.com/l/Advanced_RAG_LangGraph_Patte...

I put together four self contained notebooks exploring different ways to make a RAG pipeline more agentic. Instead of retrieving once and immediately generating an answer this methods show patterns where the system can perform the following * decide whether retrieval is needed * where to retrieve from * evaluate retrieved context * correct or retry when the context isn't sufficient * defer to a human Each notebook focuses on a different pattern and is intended to be runnable and easy to experiment with. The notebooks are free to explore on GitHub https://github.com/ChandulaSenevirathna/Agentic_RAG I also put together a more complete version with all four implementations and explanations here https://chandula7.gumroad.com/l/Advanced_RAG_LangGraph_Patte... Would be interested in feedback on which Agentic RAG pattern people are finding most