Projects
My work focuses on using mathematics, statistics, optimization, and scientific computing to understand difficult applied problems. Below are several selected projects from my academic research and consulting work.
Physiological Modeling and Regression
I currently consult with Denver Life Sciences on statistical models for physiological study data. The project uses structured and regularized regression methods to distinguish stable baseline differences among experiments from changes that occur within experiments over time.
The goal is to improve prediction and interpretation of physiological outcomes such as blood pressure, heart rate, and oxygen saturation. The work combines statistical estimation, model validation, and close collaboration with scientific domain experts.
Mean-Field Control of Autonomous Agents
This research developed scalable mathematical methods for controlling large interacting systems. Rather than optimizing the trajectory of each agent independently, the model describes the collective behavior of the population and accounts for interactions, obstacles, and control objectives.
The resulting method uses kernel approximations and numerical optimization to solve high-dimensional control problems involving thousands of simulated autonomous agents.
View the open-source implementation on GitHub
Optimal Transport and Generative Modeling
This project studied continuous normalizing flows through the perspective of optimal transport. We used the Jordan–Kinderlehrer–Otto scheme to connect the training of generative models with Wasserstein gradient flows.
The approach was designed to reduce sensitivity to difficult hyperparameter choices while providing a clearer mathematical interpretation of how probability distributions evolve during training.
Hyperspectral Mineral Modeling
Working with researchers at the Colorado School of Mines and the United States Geological Survey, I developed methods for interpreting hyperspectral measurements of geologic core samples.
The project combined image recognition, convolutional neural networks, stochastic autoencoders, and mineralogical reference data to connect measurements collected at different physical scales. The broader objective was to improve mineral mapping and support geologic interpretation.
Research and Collaboration
I am interested in projects involving mathematical modeling, optimization, inverse problems, statistical estimation, and scientific computing, particularly in the life sciences, earth sciences, physics, engineering, and environmental research.
Additional information is available on my CV and through the Google Scholar link in the sidebar. Please contact me to discuss a research or educational collaboration.