Module 7 of 10, lesson 3 of 5

Videos

Agentic RAG

Watch these first

Building Agentic RAG From Scratch in Pure Python

Agentic search with three tools: list files, grep and read, without a vector database, then what changes in production. Dave Ebbelaar, 27:28

Building Agentic RAG From Scratch in Pure Python Play video (27:28, Dave Ebbelaar)

Clicking loads the YouTube player, which may set trackers. No framework. May 2026. Code on GitHub.

• 0:48 Semantic vs agentic RAG
• 2:12 Defining the tools
• 12:37 Building the agent
• 13:31 Debugging the agent
• 20:33 Production considerations

What is Agentic RAG?

The agent picks the source, and a failsafe answers when nothing fits. IBM Technology, 5:42

What is Agentic RAG? Play video (5:42, IBM Technology)

Clicking loads the YouTube player, which may set trackers.

• 1:30 Using the LLM to decide
• 3:00 Routing between sources

If you want more

Optional, not counted in the module time.

The Complete Guide to Hybrid Search in RAG

BM25, embeddings, rank fusion and reranking, built one by one and compared by measurement. Dave Ebbelaar, 59:17

The Complete Guide to Hybrid Search in RAG Play video (59:17, Dave Ebbelaar)

Clicking loads the YouTube player, which may set trackers. Improve retrieval first. Code on GitHub.

• 10:05 BM25 keyword search
• 23:56 Why dense embeddings help
• 37:12 Reciprocal rank fusion
• 43:56 Adding a reranker
• 49:36 Evaluating retrieval

Legal notice and privacy Manage cookies