Home Research About Contact

// PhD Student

Deep learning for Dynamical Systems and Control

My interests lie at the intersection of Deep Learning, and Control Theory. Broadly, I'm interested in how learned representations can improve the way we reason about control decisions for PDE systems. Lately, I've been excited about Joint-Embedding Predictive Architectures (JEPA) and recently wrote the first paper using JEPA for PDE Control. I am really excited about this direction and foresee working on it for the remainder of my PhD. Currently a PhD student at the University of Waterloo.

JEPA Neural Operators World Modeling for PDE Physics-Informed ML Nonlinear Control PDE Surrogate modeling Uncertainty Quantification Embedded Systems

Pre-Print

Toward Goal-Agnostic Joint-Embedding Predictive Control of Partial Differential Equations

Jonathan Gallagher, Roberto Guglielmi

2026

Pre-Print

Accelerated Amine Based Carbon Capture Performance Optimization Leveraging A Fully Differentiable Neural Surrogate

Jonathan Gallagher, Roberto Guglielmi, Zhao Pan, Yunli Wang

2026
All publications →