CMU 11-768 Covers Full Agent Stack
Carnegie Mellon’s Fall 2026 AI Agents course, taught by Graham Neubig and Daniel Fried, is making lectures and course materials freely available online. The curriculum spans agent harnesses, tool use, evaluation, safety, and training models with SFT and reinforcement learning.
This is one of the rare public courses treating agents as complete systems rather than prompting demos, making it especially valuable for developers building production-grade agents.
- –Covers the full build–evaluate–train loop, including ReAct harnesses and agent-specific evaluations
- –Goes beyond coding agents into GUI agents, deep research, multi-agent interaction, and memory
- –Connects harness engineering with SFT and RL training instead of treating the model as a black box
- –Open lectures, readings, slides, and assignments lower the barrier to serious agent research
- –The graduate-level prerequisites make it better suited to experienced ML and AI developers than beginners
DISCOVERED
1h ago
2026-09-28
PUBLISHED
1h ago
2026-09-28
RELEVANCE
AUTHOR
techNmak