Stefan Broecker

Hi, I’m Stefan Broecker! I’m currently a PhD Candidate in the Department of Computer Science at UC Davis working under Dr. Thomas Strohmer in the Department of Mathematics. My PhD work sits at the intersection of research and application. I’m very interested in moving the work being done in academia into the real world. Right now, the academic fields that my work touches include large language models (LLMs), multimodal deep learning, synthetic data generation, and privacy. Some of the appications for my work include improved medical diagnosis, LLM evaluation, and privacy.

My path to get here meandered, but roughly:

• I’m currently exploring multimodal synthetic data generation, LLM unlearning, and LLM agent evaluation as part of my PhD research.
• I’m spending this summer at Okareo, where I’m focused on synthetic data generation and model evaluation for Large Language Models (LLMs). I did similar work for them last summer as well.
• My initial dive into privacy and synthetic data was with the US Census Bureau, where I built differential privacy tools for healthcare data. You can see some of that here.
• Before grad school I worked in Data Science with a particular emphasis on civic tech — an interest that continues to drive my work today.
• And prior to all of that I worked on science and technology policy as part of the University of Virginia’s Policy Internship Program.

If any or all of that sounds interesting to you, reach out! I’m always eager to collaborate on a project or talk about what’s going on in AI, privacy, or tech.