As the Lead Developer for the Willerson Center, I architect GPU-native frameworks designed to push the boundaries of computational medicine. My software focus is on differentiable physics, high-performance kernels, and Scientific Machine Learning (SciML) integration.
CARDIAX
Role: Lead Architect & Developer
Status: Public Release
Stack: JAX, Equinox
CARDIAX is a GPU-native, differentiable finite element package engineered for high-speed cardiac mechanics. By leveraging the JAX ecosystem, it enables seamless auto-differentiation through the entire FE assembly and solver pipeline.
NNFE
Role: Lead Developer
Status: SoftwareX (Under Review)
Stack: JAX, Equinox
A scientific machine learning library for learning solutions to parameterized families of PDEs. It replaces traditional constant parameters with neural networks trained directly on the variational residual of the PDE.
Research Teasers (In Development)
The following repositories are currently under active development within the Willerson Center and are slated for future open-source release.
THB-Splinax
Focus: High-Performance Isogeometric Analysis
Stack: JAX, Equinox
A dedicated kernel for Truncated Hierarchical B-splines (THB-splines). This package provides the geometric foundation for our IGA simulations, allowing for local refinement on complex manifolds while maintaining $C^k$-continuity. It is designed to bridge the gap between CAD-based topology and GPU-accelerated analysis.
Ionix
Focus: Subcellular Electrophysiology
Stack: JAX, Diffrax
A massively parallel ODE solver suite for modeling ion channel dynamics and calcium transients. Ionix is built to handle millions of independent cellular models simultaneously, providing the electrochemical “trigger” for whole-heart digital twins.
SciVis
Focus: Scientific Visualization Pipeline
Stack: PyVista
Formerly CARDIAX-Vis, this toolbox provides a high-performance bridge between GPU-resident simulation data and publication-quality visual analysis. It enables automated post-processing and headless rendering of large-scale tensor fields directly from JAX/NumPy arrays.
Open Source & Collaboration
I am a firm believer in open science. While several of my core libraries are currently in a “Research Preview” state, I welcome inquiries regarding collaboration or early access for academic use. Please reach out via the Contact Page.