AI Agents & RAG Course (Free, Python)

A free, practical course: build an AI agent that can use tools and answer questions from your own documents (retrieval-augmented generation, or RAG) — in plain Python, step by step. You need basic Python and an Anthropic API key; everything else is explained as we go.

What you will build

• A tool-using agent that decides when to call your Python functions
• A RAG pipeline: split documents, search them, and answer with sources
• A retrieval threshold so the assistant says “I don’t know” instead of guessing
• Guardrails, cost limits and logs you would want before sharing it with a team
• A finished policy Q&A assistant for your company’s documents

Lessons

1. 1What an AI Agent Really Is (and When You Don’t Need One)An AI agent is a model in a loop with tools. Learn the loop, the difference between a chatbot, a workflow and…
2. 2Your First Tool-Using AI Agent in PythonBuild a working agent loop with the Claude API: define tools as JSON schemas, run the model, execute tool calls and send…
3. 3RAG: Make AI Answer From Your Own DocumentsRetrieval-augmented generation in plain Python: split documents into chunks, score them against a question, and send only the best chunks to the…
4. 4Better Retrieval: Chunk Size, Scores and a “Don’t Know” ThresholdMeasure retrieval instead of guessing: build a small test set, compare chunk sizes, and add a score threshold so the assistant refuses…
5. 5An AI Agent That Searches Your DocumentsTurn RAG into a tool: the agent decides when to search, can search more than once with better wording, and combines document…
6. 6Guardrails, Cost Limits and Logging for AI AgentsMake an agent safe to share: turn limits, a cost ceiling, read-only tools, approval for risky actions, prompt-injection awareness and a log…
7. 7Project: Build a Policy Q&A Assistant for Your TeamPut the course together: a small web app that answers questions from your policy documents with sources, refuses when unsure, logs usage…