Experience
Leading the VLA direction for autonomous driving, post-training and distilling models on NVIDIA Cosmos backbones. Building eval infrastructure for long-tail scenario mining and closed-loop benchmarking to validate model performance before on-vehicle deployment.
Set technical vision for the PRISM subteam, defining the research roadmap for world models and differentiable simulation for physics-aware manipulation. Our work on learning object dynamics from few interactions won Best Extended Abstract at the RINO Workshop (CoRL ’25).
Trained large-scale VLAs and generative video models for robotic manipulation, bridging foundation model research and real-world deployment. Core contributor to Theia (CoRL ’24) and Real-is-Sim (ICRA ’26). Institute-wide go-to for 3D reconstruction.
Led R&D across multiple DARPA programs, shipping real-time multi-sensor perception stacks and advancing generative AI for defense. Built custom conditional diffusion models for synthetic overhead imagery (Best Paper, IH&MMSec ’23), developed diffusion-based defenses against backdoor attacks (Spotlight, ICLR ’23), and published on explainable face recognition (ECCV ’20).
Worked across computational imaging and 3D vision including real-time visual SLAM, dense 3D reconstruction, structure from motion, super resolution, and satellite image analysis. Developed a spectral band registration algorithm for the WorldView-3 satellite that was transferred to DigitalGlobe via an NGA CRADA.
Education
Imaging Science is an interdisciplinary field spanning optics, sensor design, image processing, machine learning, and computer vision. My thesis combined robotic gigapixel imaging, eye-tracking with automated gaze mapping, and semi-immersive projection to build virtual geology field trips.
Magna cum laude with honors in physics. Minor in Computer Science. Periclean Honors Society. Awarded the Eleanor A. Samworth Prize for an outstanding senior major in physics.
Interested in collaborating or just want to say hello?