I develop and study multimodal, human‑centered AI—combining computer vision, large language models, and on‑device sensing—to enable accessible navigation and context understanding for visually impaired users, and robust perception and localization in autonomous systems through human–AI collaboration.