Discover what makes code agents special and how they use Python to reason and act. Build your first agent with smolagents, add built-in and community tools for web access, and create custom tools to connect agents with data.

Introduction to Hugging Face smolagents

Let the Agent Do the Math: Expense Insights

Wait, You Didn't Write Any Functions?

Creating an Agent With Tools

Writing Code Through Prompts: Compound Interest

Let Agents Search the Web

Creating an Agent With Custom Tools

Just Add @tool: Writing a Custom Tool

Check, Please! Giving Agents Access to Data

One Prompt, Many Tools

Transform your traditional RAG pipeline into an agentic system that retrieves information iteratively and reasons across multiple steps. Build stateful tools to support advanced retrieval, guide agents with planning intervals to improve outcomes, and use callbacks to track and customize agent behavior at runtime.

RAG Refresher

The RAG Pipeline Step by Step

RAG in Order: From Manuals to Answers

So, What’s Actually Happening in RAG?

The Step That Adds Meaning

Introduction to Agentic RAG

Stateful by Design: Build a RAG Search Tool

Connect the Tool, Unlock Agentic RAG

Working With Multi-Step Agents

Track Agent Steps, Monitor the Action

Game Time: Run the Full Agent

Agentic RAG and Multi-Step Agents

Tackle complex workflows by orchestrating teams of specialized agents under a coordinating manager. Add memory to retain context across interactions, debug agent behavior using execution traces and reasoning steps, and implement robust validation strategies to ensure high-quality, reliable responses.

Introduction to Multi-Agent Systems

Divide and Conquer: Creating Specialist Agents

Meet the Team: Agent Coordination

Managing Agent Memory

Don't Forget: Keeping Memory Between Calls

Tracing the Agent’s Code Execution

Agent Final Answer Validation

Say More, Agent: Enforcing Answers Length

Smarter Agents: Validate and Retry

Congratulations!

Multi-Agent Systems, Memory and Validation

Orders

AI agents are changing how we work with data and software. From automating workflows to helping users navigate complex tasks, agents can search, reason, and act on your behalf. In this course, you’ll learn how to build agents using smolagents, a lightweight Python framework developed by Hugging Face.

Get Hands-On With Code Agents and Tools

You’ll start by understanding what makes code agents different and why they're so powerful. Then, you’ll build your first agent from scratch, using smolagents to generate and execute Python code. You’ll also learn how to plug in built-in tools and create custom tools to extend what your agents can do.

Make Agents Smarter With RAG and Memory

Next, you’ll use retrieval-augmented generation (RAG) to help agents pull info from large document collections. You’ll take things further by building agentic RAG systems—agents that reason over multiple steps to get better answers. You’ll also learn how to add memory so agents can handle follow-up questions naturally and keep track of what’s already been done.

Coordinate Multi-Agent Systems and Validate Outputs

In the final chapter, you’ll build multi-agent systems that coordinate specialist agents through a manager. You’ll add planning intervals, use callbacks for insight into agent behavior, and validate final answers, so your agents stay reliable and user-friendly.

By the end of the course, you’ll know how to build agents that think ahead, work together, and get things done.

Working with Hugging Face

Retrieval Augmented Generation (RAG) with LangChain

Learn how to build intelligent agents that reason, act, and solve real-world tasks using Python.

AI Agents with Hugging Face smolagents

Podstawy Hugging Face

Build a semantic index: chunk documents, embed, and store vectors.

Search the index and select the top most relevant chunks.

Assemble context: join retrieved chunks with separators and metadata.

Generate answer: send question and context to the LLM for a response.

RAG in Order: From Manuals to Answers

AI Agents with Hugging Face smolagents

Interaktywne ćwiczenie praktyczne