The evolution of Large Language Models (LLMs) has shifted from simple text generation to the "agentic" paradigm—where AI doesn't just describe solutions but actively executes them. However, for AI to tackle quantitative domains including automotive engineering it requires more than just language; specialized toolkits are needed to navigate complex 3D environments and satisfy rigorous physics constraints. In this talk, I introduce Datum, a San Francisco-based startup building the bridge between agentic AI and physical engineering. I will explore how we enable AI agents to search 3D geometries and automate physics simulations, putting decades of institutional knowledge and computational rigor at every engineer’s fingertips. To conclude, I will share reflections on the current job market and search process, offering a founder's perspective on what it means to build—and find—a career at the frontier of this new technology.

BIO:
Kevin Nelson is a Founding Engineer at Datum Systems Inc, a San Francisco-based startup specializing in AI agents for engineering design.

Before Datum, Kevin spent 11 years in particle physics research, starting at the College of William and Mary before coming to Michigan for his PhD and postdoctoral research with the ATLAS experiment. His research applied geometric deep learning to beyond the standard model searches in the Higgs sector. Now, at Datum he applies the same techniques to artificial intelligence. When not working on AI, you can usually find him out for a run in the beautiful San Francisco weather.