Ph.D. Student Samuel Rothfarb Wins National Award for AI-Driven Materials Research

 

Image of Samuel Rothfarb (left) and Wilton Kort-Kamp (right) standing in front of a research poster titled “LLM Agents for Autonomous Density Functional Theory Simulations from Natural Language.” The poster displays diagrams and figures related to quantum simulations and AI agents. Both individuals are smiling, and the setting appears to be a conference poster session.

Samuel Rothfarb (left) and Wilton Kort-Kamp (right) standing in front of their poster

Image of Samuel Rothfarb receiving the NVIDIA GPU Award during an outdoor ceremony. He is smiling and holding a certificate while standing next to an event organizer. Three other attendees are standing nearby, applauding.

Samuel Rothfarb receiving the award

 

Samuel Rothfarb, a first-year Ph.D. student in the School of Civil and Environmental Engineering at the University of Connecticut and advisee of Professor Baikun Li, has been awarded the prestigious NVIDIA GPU Award for Outstanding Research and Poster Presentation at the 2025 Machine Learning in Chemical and Materials Science (MLCM) Conference in Santa Fe, New Mexico. The event was co-sponsored by Los Alamos National Laboratory (LANL), Software for Chemistry & Materials, and NVIDIA.

Rothfarb’s winning poster, “LLM Agents for Autonomous Density Functional Simulations from Natural Language,” presented cutting-edge work that explores how multi-agent large language model (LLM) systems can autonomously run quantum simulations and design new materials. The project, done in close collaboration with LANL researchers Edward Holby and Wilton Kort-Kamp, integrates OpenAI’s LLMs with NVIDIA’s CUDA-powered GPUs to streamline high-throughput density functional theory (DFT) simulations, which removes barriers of fragility, inefficiency, and technical complexity that plague traditional automated workflows.

His work utilizes “agentic” systems, in which specialized LLMs are deployed with tailored roles, such as error correction, input generation, or planning, allowing for flexible, real-time decision-making during simulations. This enables scientific exploration at a level of autonomy and adaptability not previously possible, especially in applications such as catalyst discovery and quantum materials modeling.

Competing against approximately 30 researchers – many of them postdoctoral scholars from elite institutions such as UCLA, Carnegie Mellon, and Oak Ridge National Laboratory – Rothfarb was the only participant from environmental engineering. He received the top prize and an NVIDIA RTX 4000 Ada GPU to support future computational work. 

“As both an undergraduate and now a Ph.D. student in the SoCEE, I’ve experienced firsthand how the school pushes students to pursue ambitious goals and innovative ideas,” Rothfarb shared. “Receiving this award feels particularly meaningful knowing it builds on the strong foundation I gained under the School’s guidance.”

This award builds on Rothfarb’s growing involvement with LANL. He was selected for a competitive summer internship through a UConn–LANL collaboration spearheaded by Dr. Alexander Balatsky of UConn Physics. One of only two UConn graduate students selected in 2025, Rothfarb is now an Institute of Materials Science Fellow in LANL’s Theoretical Division 4 (Quantum and Condensed Matter Physics), where he continues advanced research on LLM agents – now applied to energetic and critical materials. His work has already garnered attention from senior directors at LANL and from leaders at NVIDIA, including Justin Smith, Principal Manager of AI for Chemistry and Materials.

This achievement highlights Rothfarb’s emerging leadership in scientific AI innovation.

MLCM 2025 Poster