January 20, 2026 by Marni Ellery

UC Berkeley professors across four engineering departments have been awarded funding to integrate artificial intelligence into their undergraduate courses. Through a generous gift from the Center for Advancing Women in Technology (CAWIT) and a donor-supported campus initiative, faculty will broaden the exposure of students to AI tools.

“The focus of this initiative is to encourage novel integration of AI tools into the teaching of core application areas in engineering, as opposed to the creation of courses on AI theory and methods,” said Clark Nguyen, executive associate dean, UC Berkeley College of Engineering. “These pilot projects reflect Berkeley’s commitment to providing access to an excellent undergraduate education in a rapidly changing world.”

Faculty whose proposals received funding will spend time developing their ideas over the spring 2026 semester and the following summer, with the goal of implementing them by the 2026–27 academic year. After the conclusion of the course, they will evaluate how well their approach enhanced the learning experience for students.

“I am grateful to our faculty for lending their vision and expertise to explore the use of AI for engineering education,” said Mark Asta, interim dean of engineering. “These instructional enhancements are just one more way that Berkeley Engineering is driving innovation in advancing the future of undergraduate education.”

The following proposals will receive funding through an initiative led by Berkeley Engineering and supported by CAWIT:

Cultivating the Next Generation of Engineers in the Age of AI: Prompting, Judging, Partnering and Planning Smarter

Huiwen Jia, assistant professor of industrial engineering and operations research, aims to help students work with AI as a thinking partner, an approach she believes is important as the nature of entry-level work evolves. Jia will integrate AI into two core IEOR undergraduate courses, giving students opportunities to apply AI in real-world modeling contexts. Students will learn to craft prompts, critique AI-generated results and integrate those outputs into rigorous modeling frameworks, so that AI becomes part of the analytical workflow. This approach is intended to cultivate a learning space where students can experiment, critique and refine their use of AI — preparing them for long-term engagement with AI-assisted decision-making.

CalorAI: A Scaffolded Small Language Model for Resource-Grounded Guidance in Engineering Thermodynamics Education

Thomas Schutzius, associate professor of mechanical engineering, and Claudio Hail, assistant professor of mechanical engineering, will develop a language model-powered assistant to provide targeted guidance for thermodynamic problem solving. When confronted with a homework problem, students can input their query and receive intelligent suggestions for specific preparatory materials, rather than a problem solution. The “prompts/completions” are scaffolded, allowing students with a wide range of understanding to engage with the model and receive meaningful feedback on their approach. The assistant can direct students to particular book chapter sections, illustrative exercise problems, relevant segments of lecture videos, and sections within solved Jupyter notebooks from in-class examples. Schutzius and Hail anticipate fine-tuning the model each semester to keep it relevant.

PupilBot — Advancing Engineering Education Through the Protégé Effect

Daniel Pirutinsky, assistant teaching professor of industrial engineering and operations research, has developed PupilBot to harness the protégé effect, whereby students learn by teaching. PupilBot is a Gemini-based custom agent that acts like a novice student confused about the subject matter. Following an exam, students can earn partial credit for problems they answered incorrectly by successfully teaching the concepts to PupilBot. By challenging students to “teach” core engineering concepts, PupilBot compels them to deconstruct complex topics, identify gaps in their own knowledge and articulate their understanding with precision. According to Pirutinsky, this process can foster a more robust and lasting conceptual mastery than traditional study methods.

Integrating AI in CEE Undergraduate Education

Mohamad Hallal, assistant teaching professor, and Luis Ceferino, assistant professor, both in the Department of Civil and Environmental Engineering, are focused on embedding AI tools into core subdisciplines — from structural and environmental modeling to systems and transportation design. Their initiative, dubbed AI4CEE, emphasizes effective and responsible engagement with AI tools: prompting, verifying and interpreting outputs while recognizing their limitations and ethical implications. This project will develop and pilot short AI literacy modules that instructors can integrate into existing courses. An industry–faculty workshop will complement this effort by identifying the competencies most essential for the future engineering workforce.

The following proposal will be funded with a grant through Reimagining Higher Education, a campus-based initiative:

Askademia: Transforming Lecture Engagement in EECS Courses

Narges Norouzi and John DeNero, associate teaching professors of electrical engineering and computer sciences, have developed Askademia, a context-aware system that responds to student questions in real-time during both live lectures and asynchronous viewing. To deliver fast, accurate answers, Askademia draws from the latest instructor audio and slide text, as well as prior lecture content, course notes and discussion forums. Askademia was deployed in CS 189 and Data 100 for the fall 2025 semester, where it responded to an average of 1,000 synchronous and asynchronous student questions each week. Next, Norouzi and DeNero aim to extend Askademia’s deployment across gateway and high-demand EECS courses, potentially reaching 8,000 students per semester.