AI Agents Course
Hugging Face
Explains how agents work (tools, the think, act and observe loop, chat templates) and then builds them with smolagents, LangGraph and LlamaIndex. Exercises run in Hugging Face Spaces. In the final unit you build an agent for a benchmark and see its score on a class leaderboard. Bonus units cover fine-tuning for function calling, observability, and an agent that plays Pokémon battles. Basic Python is enough to start, and the certificates are free.
What you'll learn
- How agents reason, call tools and act on observations
- Building agents with smolagents
- Agent workflows with LangGraph and LlamaIndex
- Fine-tuning a model for function calling
- Evaluating and observing agents
- Building an agent for a benchmark
