Aryan Keskar

Worked on computer vision and multimodal reasoning research with applications in autonomous systems. Contributed to research published in a WACV 2025 workshop (Best Paper Award), focusing on improving scene understanding and hazard detection using vision-language models. - Implemented and evaluated multimodal models (e.g., VILA, OmniVLM, GPT-4o-mini) for visual question answering on driving datasets - Designed and tested prompt strategies to enhance performance on safety-critical scenarios. - Contributed to the development of novel benchmarks for hazard detection and open-set evaluation

Excited to join the Master of Science in Computer Science program at the University of California, San Diego starting in September 2026. Focused on areas including Machine learning, Computer Vision and Ariticial Intelligence. Actively preparing through independent study, projects, and relevant coursework while anticipating the start of the program. This opportunity reflects my dedication to advancing my technical expertise and contributing to cutting-edge research and development in computing.