Announcement Detail

Wednesday, September 2, 2026

11:00 AM PDT

Join via Zoom: https://us06web.zoom.us/j/81593085776?pwd=c5k00ypHOLXZbFnygG8HrbbMYcifLL.1

Student Chapter Seminar Series

Generative AI for Materials Discovery: Using Food as a Model System

Speaker

Vahidullah Taç, Stanford University

Abstract

What can the perfect burger teach us about discovering the next generation of materials?

Designing new materials often means searching through an enormous number of possible combinations while balancing competing goals such as performance, sustainability, and cost. Generative AI is emerging as a powerful tool for tackling this challenge by learning from existing successful designs and proposing entirely new ones.

Testing these methods directly on advanced materials, however, is often slow and expensive. Instead, Dr. Vahid Tac turns to an unexpected but surprisingly powerful model system: *burgers*. Burger recipes form a rich combinatorial design space in which ingredient choices and proportions determine taste, nutrition, and environmental impact—making them an ideal experimental playground for generative design algorithms.

In this talk, Dr. Tac will show how generative AI can rediscover classic recipes, invent entirely new burgers optimized for different objectives, and even learn human preferences through blinded restaurant taste tests. Along the way, he will illustrate how the same principles can be applied to accelerate the discovery of advanced materials and navigate complex design trade-offs.

Bio

*Dr. Vahid Tac* is a Schmidt Science Postdoctoral Fellow in Prof. Ellen Kuhl's lab at Stanford University. He received his PhD in Mechanical Engineering from Purdue University, where his research focused on physics-constrained machine learning, constitutive modeling of soft tissues, inverse problems in mechanics, and generative models for hyperelastic materials. As a Schmidt Science Fellow, his current research explores how generative AI can accelerate materials discovery, using food as an experimentally accessible model system to develop methods with broad applications in engineering and materials science.