AI agents are rapidly emerging as a practical area within generative AI. If you are just starting, I recommend first understanding how agents reason, use tools, retrieve information, plan tasks, evaluate outputs, and interact with external systems. Once you have this foundation, you can explore frameworks such as LangGraph, LlamaIndex, smolagents, or Microsoft Foundry. I have reviewed the current free learning options and selected four courses that are particularly useful for those interested in building AI agents rather than only learning terminology.
Free Courses to Learn AI Agents
Below are four free courses to help you learn how to build AI agents, not just understand the terminology.
1. Hugging Face AI Agents Course
The Hugging Face AI Agents Course is my top recommendation for those serious about learning to build agents.
The course begins with fundamentals such as defining agents, understanding tool use, the thought-action-observation cycle, and agent-environment interactions. It then covers practical frameworks including smolagents, LlamaIndex, and LangGraph.
It also addresses agentic RAG, includes a final project focused on building and evaluating an agent, and offers units on function calling, observability, and evaluation.
2. DeepLearning.AI — Agentic AI
Andrew Ng’s Agentic AI course emphasizes the underlying patterns of effective agentic systems rather than focusing solely on frameworks.
The course covers reflection, tool use, planning, and multi-agent workflows, as well as connecting agents to databases, APIs, web search, and code execution. It also emphasizes evaluation, error analysis, latency, cost optimization, and production deployment.
Notably, the course uses Python and teaches many concepts from first principles before introducing frameworks.
Preparing for AI/ML Interviews?
If you are learning AI agents to pursue a career in AI, I recommend preparing for interviews as you progress. Cracking Your First AI/ML Interview is a valuable resource for those preparing for their first AI or machine learning interview.
3. Microsoft Learn — Develop AI Agents on Azure
Microsoft Learn offers a free Develop AI Agents on Azure learning path that covers agent development with Microsoft Foundry Agent Service and Microsoft Agent Framework. This intermediate path takes about 10 hours to complete.
The curriculum covers building, testing, deploying, and integrating agents, as well as extending them with tools.
Microsoft also offers a shorter beginner learning path on AI applications and agents if you are not ready for the intermediate material.
4. DeepLearning.AI — Evaluating AI Agents
A common mistake when learning about agents is focusing only on building them and neglecting evaluation.
The free Evaluating AI Agents course from DeepLearning.AI and Arize addresses this issue. It covers observability, tracing, component-level evaluations, LLM-as-a-judge, trajectory evaluation, experiments, and production monitoring.
You will build an agent and progressively add tracing and evaluation features.
The Takeaway
Here are four free courses to help you learn to build AI agents, not just understand the terminology:
After learning the fundamentals, begin building projects. For example, create a coding agent, then a research agent, followed by an agentic RAG system, and finally a multi-agent workflow. This is where the concepts become practical.
I hope you found this article on four free courses to learn AI agents helpful. For more AI and machine learning tips, feel free to follow me on Instagram.
