About Me

I am a final year PhD student in the EECS department at MIT advised by Professor Gregory Wornell. I received my M.S. degree in EECS from MIT in January 2022 and my B.S. degree in Electrical Engineering from the University of Illinois at Urbana-Champaign in 2019. My research interests lie at the intersection of generative modeling and representation learning. My current interests are in developing new techniques for score-based generative modeling such as improved techniques for training diffusion models and using these models for designing new one-step generators for downstream tasks such as inverse problems. I am also generally interested in the interplay between information theory and representation learning and have worked on several projects related to neural compression with generative decoders such as vision transformers, GANs and diffusion models.

Research Internship Experience

I have had the opportunity to complete multiple research internships in the past. Currently, I am wrapping up my internship at Adobe Research where I am working on one-step generative modeling by leveraging novel diffusion distillation techniques that enforce distribution similarity. Before this, I worked on transformer-based video compression architectures with Dr. Fabian Mentzer on the Neural Compression Team at Google Research (see slides). I also worked with Dr. Qing He and Dr. Vimal Manohar at Meta AI in 2021 and 2022 respectively where I worked on speech compression and singing voice conversion with generative models. I have also interned with Dr. Jonathan Le Roux at MERL in 2019, where I conducted research on adversarial attack detection.