About

I am an applied mathematician interested in the interplay of scientific computing and artificial intelligence. I am a Winship Distinguished Research Associate Professor in the Department of Mathematics and the Department of Computer Science at Emory University and a member of Emory’s Scientific Computing Group. I lead the Emory REU/RET site for Computational Mathematics for Data Science. Prior to joining Emory, I was a postdoc at the University of British Columbia and I held PhD positions at the University of Lübeck and the University of Münster.

Research

Scientific Computing Numerical algorithms for high-dimensional differential equations, optimization, and inference
Artificial Intelligence Generative models, continuous-time deep learning, mixed-precision training, and efficient numerical optimization

My research lies at the intersection of computational mathematics and artificial intelligence. I develop algorithms that make AI models more efficient, stable, and interpretable, and I use AI to address challenging problems in numerical modeling, inference, and control. My methods are rooted in numerical algorithms for differential equations (ODEs, SDEs, PDEs) and tools from numerical analysis, high-performance computing, optimal transport, and optimal control.

As a scientific computing expert working in AI, I am interested in continuous-time deep learning, treating neural networks as dynamical systems that can be analyzed and trained with established numerical methods. I am also interested in optimal-transport–based generative models, faster optimizers that exploit problem structures such as separability, leaner architectures, and mixed-precision algorithms for training quantized networks.

As an AI/ML researcher working on computational mathematics, I am interested in using neural networks to approximate value functions and transport maps, enabling high-dimensional optimal control, mean field games, and Bayesian inverse problems while incorporating structure from HJB equations and the Pontryagin Maximum Principle. I am also interested in learnable iterative solvers to accelerate challenging PDE simulators.

My research is collaborative and I have experience working with national laboratories and industry partners. I am always open to new collaborations, so please feel free to reach out if you are interested in working together!

Teaching

I regularly teach undergraduate and graduate courses at all levels from introductory mathematics courses to special topics seminars in computer science and mathematics. My teaching bridges foundational techniques from numerical analysis, optimization, and differential equations and modern applications in artificial intelligence and other fiels. I use flipped classroom models in all my classes to provide experiential learning opportunities to students to help them become effective life-long learners.
Deep Generative Modeling Workshop Interactive three-hour mini-course covering normalizing flows, variational autoencoders, and generative adversarial networks with hands-on Python examples
Math 785R: Deep Generative Modeling Mathematical foundations of deep generative models, emphasizing theoretical principles and connections to optimal transport, high-dimensional probability, and dynamical systems

Selected Service

Team

I enjoy mentoring highly motivated students and early career researchers. I frequently advise Emory Undergraduates in our Honors program and undergraduate students from other US institutions at our REU site. I also mentor Ph.D. students in Emory’s Computational Mathematics and Computer Science and Informatics Ph. D. programs.

Current Group Members

  • Katie Keegan (Computational Mathematics PhD student)
  • Rishi Leburu (Undergraduate Honors student)
  • Haley Rosso (Computational Mathematics PhD student, main advisor: Talea Mayo)
  • Warin Watson (Computational Mathematics PhD student)

Current Projects

Scalable Graph Neural Network Algorithms and Applications to PDEs We combine geometric machine learning and numerical PDEs and seek to improve the scalability of graph neural networks and enhance PDE solvers.
Collaboration with Eran Treister's group. Funded by NSF/BSF
2024-2027
Mixed-Precision Algorithms for Deep Learning We develop mixed-precision algorithms to accelerate the training of deep neural networks for applications in science and engineering.
Funded by ONR
2024-2027
REU Site: Computational Mathematics for Data Science Our REU site offers 8-week summer research opportunities at Emory.
Funded by NSF
2023-2026
RTG: Computational Mathematics for Data Science Our RTG develops mathematical theory, models, and computational algorithms for data science with a broad range of applications.
Funded by NSF
2021-2026