I am Sudhir, a machine learning research scientist at Adobe. I completed my PhD at the University at Buffalo under the supervision of Prof. Yuan Junsong. During my PhD, my research focused on designing video perception and forecasting models for autonomous systems, enabling them to perceive and forecast in the real world just as humans do, and ultimately surpass human forecasting abilities. My research spans two core directions: Video Perception, where I develop efficient and domain-adaptable models; and Future Forecasting, leveraging video perception to accurately predict future scenes.

At Adobe, my research interests lie at the intersection of computer vision and machine learning, leveraging insights from behavioral responses of people to multimodal content. I focus on building foundational generative models in the video domain and their applications in marketing and advertising scenarios.

Feel free to contact me by email if you want to discuss this further: ykumar@adobe.com. To get more information, check outmy page at https://skrya.github.io/.