Jay Parmar

I’m a Computer Science PhD student at the University of Central Florida working in the Center for Research in Computer Vision (CRCV). My research centers on vision-language models (VLMs), video understanding, and fine-grained retrieval, with a special focus on addressing privacy preservation and bias mitigation.

I've co-developed a benchmark dataset for Composed Video Retrieval (CoVR), and contributed methods that outperform prior models on tasks like action recognition and video-text alignment. My recent work includes anonymization-aware VLMs for pedestrian and traffic footage, and large-scale YouTube data collection for safety systems.

Previously, I conducted machine learning research on solar panel diagnostics using IV curve classification and image-based fault detection. I also led and mentored undergrad students in CS fundamentals as a Supplemental Instruction Leader, receiving recognition for my teaching support.

I enjoy building responsible, scalable AI that serves real-world needs, with open-source contributions and public benchmarks as a key part of my philosophy.