Research

Scientific computing, machine learning, visualization, and responsible AI
My current research brings together scientific computing, machine learning, visualization, and responsible AI to develop systems that are rigorous, understandable, and useful in the world. Across four connected areas, I pair foundational methods with collaborations that address accountability, equity, and complex public and engineering decisions.
Current research is organized around four connected directions.
Research Direction I

Responsible AI and Human-Centered Systems

With philosopher David Danks, I am developing ways to reason about accountability, governance, and human responsibility in AI-mediated systems. I also study how AI can support critical engagement, literacy, and responsible design in organizational settings.
  • Accountability and governance for AI-mediated systems
  • AI literacy, digital leadership, and participatory design
  • Human and model interpretation in online information environments
Research Direction II

AI for Mobility and Public Decision-Making

With Guang Tian, UDOT, and UTA partners, I am integrating transit-agency data with aggregated, anonymized mobility data to understand travel patterns and service gaps. I use interpretable AI and machine learning to explore route, frequency, and multimodal improvements that can inform public-agency decisions.
  • Interpretable models for mobility demand and service gaps
  • Responsible use of integrated mobility data
  • Decision support for equitable public services
Research Direction III

Scientific Machine Learning, Digital Twins, and Engineering AI

I develop physics-informed and scientific machine-learning methods that incorporate governing equations, physical constraints, and data. My work builds multifidelity surrogates, neural operators, and Gaussian-process approaches for efficient simulation, optimization, and digital-twin applications.
  • Physics-informed learning for differential equations
  • Multifidelity and derivative-informed surrogate modeling
  • Reliable and interpretable AI for engineering systems
Research Direction IV

High-Order Methods and Scientific Computing

I advance high-order finite and spectral/hp element methods for accurate, efficient solution of complex partial differential equations. This work includes stable filtering, compact SIAC reconstruction, positivity-preserving mappings, and fast evaluation techniques for high-order simulations.
  • High-order algorithms and the Nektar++ ecosystem
  • Discontinuous Galerkin methods and SIAC reconstruction
  • Scientific visualization, verification, and scalable computation
These directions are deliberately connected: trustworthy AI for complex systems requires sound mathematical models, interpretable learning methods, careful data practices, and clear communication of uncertainty. See the publications page for representative work.
Last Updated: 10/12/26