Summary
Tom Bertalan is a founder and applied data scientist with 13 years of experience marrying machine learning, mechanistic simulation, and control for regulated bioprocessing and robotics. After a PhD at Princeton and postdocs at MIT and Johns Hopkins, he built production-grade digital twins and hybrid mechanistic/black-box models at Amgen, specializing in lyophilization, fill‑finish, and CHO metabolism. He co-designs data architectures and pipelines for high‑volume sensor ingest and has shipped containerized, versioned modeling packages and real‑time state estimation used company‑wide. His research blends neural differential equations, symbolic regression, and Bayesian soft‑sensing—work published in Nature Communications, PNAS, and Chaos—and he’s now applying those ideas to a stealth AI platform targeting regulated industries. Notably, he combines grammatical-evolution symbolic programming with differentiable simulators to make models that are both interpretable and deployable in real-time control.
13 years of coding experience
5 years of employment as a software developer
Bachelors of Science, Chemical and Biological Engineering, Bachelors of Science, Chemical and Biological Engineering at The University of Alabama
Doctor of Philosophy (PhD), Chemical and Biological Engineering, Doctor of Philosophy (PhD), Chemical and Biological Engineering at Princeton University