About
I grew up in Tuscaloosa, Alabama, where in my senior year of high school I developed a deep interest in physics and astronomy through pop-sci offerings like Neil deGrasse Tyson's Cosmos and Brian Greene's The Elegant Universe. I completed my undergraduate studies at the University of Virginia, double majoring in Physics and Engineering Science. While there, I worked with Prof. Craig Group building cosmic ray veto panels for the Mu2e experiment, and with Prof. Chris Neu on projects tied to the CERN Compact Muon Solenoid (CMS) detector.
I continued my CMS work as a PhD student under Prof. Dan Marlow at Princeton University, culminating in a dissertation focused on a search for a pseudoscalar Higgs boson decaying to Zh with ℓℓ + ττ final states. During my PhD years I developed a strong interest in machine learning, particularly charged particle track reconstruction via graph neural networks. That led to collaborations with Boris Hanin, where we characterized and alleviated numerical issues like poor initialization, oversmoothing, and correlation collapse in graph neural networks (GNNs).
My postdoc is focused on furthering the study of numerically stable GNNs—especially graph transformers—by tuning them for hyperparameter transfer: models are parametrized so they all agree on the same learning rate. I'm now working on several scientific applications for these models, including stellar stream identification in large astronomical surveys and jet tagging in collider datasets.
Selected papers
Heavy pseudoscalar Higgs search (A → Zh → ℓℓττ)
Search for a pseudoscalar Higgs boson A decaying to Zh with ℓℓττ final states in the full CMS Run 2 dataset.
Read more → Nature Reviews · 2023Graph Neural Networks at the Large Hadron Collider
A survey of graph neural networks and the LHC physics tasks they enable, written for both physicists and data scientists.
Read more → SIMODS · 2025Principles for Initialization and Architecture Selection in GNNs
Principles for developing numerically stable GNNs that train to arbitrary depth and width, with full characterization of initialization and smoothing control.
Read more →Selected tutorials
Creative side quests
Outside of academia, I'm an avid progressive metal guitarist, writing and recording music with several bands. You can explore recent riffs, performances, and works-in-progress on my music profile.
Let's collaborate
Available for conversations on graph based ML, detector reconstruction, and inclusive physics education.