Stanford CS329Z

Stanford CS329Z teaches students how to engineer agentic systems, from simple LLM pipelines to compound AI systems and autonomous agents. The course covers core components like RAG, tool use, and agent loops, along with frameworks such as DSPy, through hands-on assignments and a quarter-long project. Students learn to decompose problems, select components, collect data, build evaluations, and reason about design tradeoffs in building these systems.