CMU Launches Free YouTube Course on AI Agents

CMU is publishing a free YouTube course, 11-768, on AI agents, covering everything from agent scaffolds to model training, with the first assignment already available on GitHub.
CMU is releasing a new course on AI agents — 11-768 — for free on YouTube. This fall it is taught by Graham Neubig and Daniel Fried. This is one of the first courses about agents that covers the entire path from building an agent scaffold to training the model it runs on.
The course starts with expected topics:
- tool use;
- context management;
- skills;
- memory;
- planning.
It then has separate sections on coding agents, GUI agents, and deep research agents. Later topics include supervised fine-tuning, reinforcement learning, training systems, sandboxes, credential management, OpenHands, LangGraph, observability, agent-to-agent and agent-to-human interaction, and search.
The assignments best describe the course:
- Build an agent scaffold.
- Design tests to evaluate it.
- Train an agent with reinforcement learning.
The first assignment and starter code are already published on GitHub.
Lectures are appearing on YouTube throughout the semester. Graham Neubig said all videos will be included in a public playlist.
This is far more useful than learning one agent framework and assuming you now understand agents. The course shows the whole system: the model, the scaffold around it, tools and memory, evaluation, and training that can change the model itself.
Why it matters
AnalysisThe course treats AI agents as a full system — model, scaffold, tools, memory, evaluation, and training — rather than a single framework tutorial.
Discuss in community
Share your questions and insights with developers
Related in Ecosystem
Practical guides, core terminology, and agent skills
CS329A: Self-Improving AI Agents. Stanford University Course
A Stanford course gathers lectures and materials on self-improving AI agents, along with a collection of recent scientific publications.
Codync: open-source Rust tool turns AI agents into chatbots
Codync is a free Rust project that wraps more than 40 local AI agents into chatbots, enabling control from iPhone, Mac, Windows, Linux, or the terminal.
Jev and System One AI: A Complete Practical Guide to the Decision Model from TypeSafe
A deep engineering breakdown of the flagship Jev model: System One architecture versus LLMs, Choice/Score/Noul primitives, request parallelism, confidence mathematics, video use cases, code, and an overview of local open-weight alternatives (Laya, GLiNER2.5, CLM-8B).
Claude Code Projects: How a Single AI Orchestrates a Team of Autonomous Agents
A deep dive into Claude Code Projects: Coordinator and Threads architecture, parallel branches, shared memory via MEMORY.md, and cloud agent management.