Published several papers including 'DynaGuard: A Dynamic Guardrail Model With User-Defined Policies' (In Submission), 'Dense Backpropagation Improves Routing for Sparsely-Gated Mixture-of-Experts' (NeurIPS 2025), and 'Video Diffusion Models Encode Motion in Early Timesteps' (In Submission).
Research Experience
Interned at the Center for Human-Compatible AI at UC Berkeley, working with Raj Movva and Emma Pierson; conducted academic research under Dr. Pete Kyle at the University of Maryland.
Education
PhD: New York University, advised by Pavel Izmailov and Andrew Gordon Wilson; Bachelor's Degree: University of Maryland, mentored by Abhinav Shrivastava, Ashwinee Panda, and Tom Goldstein.
Background
PhD student at New York University, with research interests in machine learning and artificial intelligence. Graduated from the University of Maryland with a bachelor's degree and worked in multiple labs.
Miscellany
Occasionally posts essays on his blog as an exercise to write more often and get used to publishing imperfect work publicly.