Multimodal models for perception.

(Google's self-driving car project)

Vision-language perception and language-driven behavior planning for open-world autonomy.

Level-4 autonomous driving stack for a custom platform: HD mapping, GPS localization, 3D perception (detection + tracking), map-conditioned prediction, and behavior/motion planning.

• Top-tier award - USDOT Intersection Safety Challenge 2024

• Spot Recognition Award

Dynamic 3D scene prediction using Neural Radiance Fields (NeRFs) - GRASP Lab

CIS 581 - Computer Vision and Computational Photography (Spring 2023)

CIS 580 - Machine Perception (Fall 2022)