Harsh Kumar
I am a 2nd year PhD student working with Dr. Joseph Jay Williams. My primary research focus is on using reinforcement learning algorithms (like contextual bandits) to optimize the interaction of users with technology. As part of Microsoft's AI for Accessibility grant, I have been working on an adaptive text-messaging system to personalize the content and delivery of text messages for the mental well-being of users, based on the contextual information of the users. Another line of work focuses on how pre-trained general AI models could augment user workflows. I aim to bridge my research in adaptive experiments with pre-trained language model research to explore the design space of self-improving and personalized conversational agents to improve the everyday lives of users. I will be interning at Microsoft Research in the Computational Social Science group this summer.