Stanford CS25: Transformers United V6 I Overview of Transformers

Why It Matters

Understanding transformers equips students with skills essential for AI development, while the industry partnership accelerates talent pipelines and innovation in a sector dominated by large‑scale models.

Key Takeaways

• Course introduces Transformers via expert speakers from academia and industry.
• Attendance required; lectures recorded, posted on YouTube after weeks.
• Sponsors Modal AI House and MongoDB offer networking and $1,000 project prize.
• Curriculum covers evolution from hand‑engineered features to self‑supervised Transformers.
• Emphasis on practical applications, limitations, and open research problems.

Summary

Stanford’s CS 25 “Transformers United” returns for its sixth iteration, a graduate‑level seminar that demystifies the transformer architecture that underpins today’s AI boom. Co‑instructors Steven, a fourth‑year CS PhD, and Karan, a third‑year EE PhD, frame the course as a bridge between cutting‑edge research and student‑level mastery.

The syllabus blends weekly expert talks—spanning computer vision, natural language processing, biology and neuroscience—with a mandatory attendance policy. Lectures run Thursdays, both in‑person and via Zoom, are recorded and posted to YouTube after two weeks. Sponsors Modal AI House and MongoDB fund a Frontier Lunch Club, offering direct access to founders, a $1,000 project prize, and a recruiting pipeline.

In the introductory session the instructors trace the field’s evolution from hand‑engineered features to self‑supervised transformers, illustrating concepts such as masked autoencoders, word embeddings, and multi‑head self‑attention with a library‑search analogy. They highlight why recurrent networks have been eclipsed—parallelism, longer context windows, and scalable training—while noting open challenges like optimal masking strategies and interpretability.

For participants, the course promises not only a solid technical foundation but also networking leverage in a rapidly expanding AI talent market. By exposing students to frontier research and real‑world deployment scenarios, it cultivates the next generation of contributors who can address the scalability, efficiency, and ethical questions that accompany ever‑larger transformer models.

Original Description

For more information about Stanford’s graduate programs, visit: https://online.stanford.edu/graduate-education

April 2, 2026

This seminar covers:

• Overview of the history of ML/NLP, Transformers, and how they work

• Recent trends, breakthroughs, applications, and current challenges

Follow along with the seminar schedule. Visit: https://web.stanford.edu/class/cs25/

Instructors:

• Steven Feng, Stanford Computer Science PhD student and NSERC PGS-D scholar

• Karan P. Singh, Electrical Engineering PhD student and NSF Graduate Research Fellow in the Stanford Translational AI Lab

• Michael C. Frank, Benjamin Scott Crocker Professor of Human Biology Director, Symbolic Systems Program

• Christopher Manning, Thomas M. Siebel Professor in Machine Learning, Professor of Linguistics and of Computer Science, Co-Founder and Senior Fellow of the Stanford Institute for Human-Centered Artificial Intelligence (HAI)