This project-based course will introduce students to building and evaluating agentic AI applications powered by foundation models. The overriding theme of the course is that building an initial prototype AI system can often be done easily but refining a prototype into something that is useful and reliable requires iterative improvement based on clear evaluation metrics. We will cover background in foundation models, prompting, and retrieval-augmented generation (RAG) before introducing full agentic AI architectures. For each architecture, the course will study methods for evaluation. Students will complete introductory homework assignments to become familiar with retrieval-augmented generation (RAG) and agentic AI. Students will then work in pairs or small teams to develop applications using agentic or other approaches and evaluate them by adapting evaluation methods presented in the class.
Introductory Python-based ML class (equivalent to CS229), knowledge of deep learning (such as CS230, CS231N, etc.), and familiarity with ML frameworks in Python like PyTorch, Keras, or TensorFlow.
Permissive but strict. If unsure, please ask the course staff!
We are generally open to in-person auditing requests by Stanford affiliates. Audits can attend lectures if there are available seats after enrolled students have taken their seats. Please come to the lecture, and if there is space available, audits may stay for the session.
Audits are welcome to participate in the discussion in lectures, but we encourage that audits check out the slides from previous lectures first — lectures could be confusing otherwise! Audits are also welcome to attend OHs, though if there are more enrolled students waiting in the line, they’ll be given priority.
We are unable to support remote audits as lectures are held in-person, and are not recorded. The general public can follow along with the course by accessing the published lecture slides, notes, assignments, and final report instructions online. These materials are available on the Syllabus page.
Please note that no course support will be available for auditors. However, we encourage you to share any feedback on the course — if there’s any concept that you find confusing, incorrect or missing, please let us know!
There's no textbook. The course relies on lecture slides and accompanying readings.
The quarter long course project has five major deliverables:
