Harshit Varma | हर्षित वर्मा

Hi, I am a final-year undergraduate at IIT Bombay pursuing a bachelor’s degree (with honours) in Computer Science and Engineering. My current research interests lie (broadly) in machine learning. I am particularly interested in the foundational aspects of deep learning, such as optimization, representation learning, different kinds of supervision, improving the generalizability and robustness of deep neural networks, adapting them to unseen domains, incorporating task-specific constraints, and other related topics.

I am currently interning with Abacus.AI Research, working on LLMs. I am also working with Prof. Sunita Sarawagi on domain adaptation methods for semantic parsing and with Prof. Suyash Awate on robust medical image segmentation. Last summer, I interned at Adobe Research, where I worked on the extraction of event-event relations from natural language text for document understanding. In the past, I have worked with Prof. Thomas Deserno at Technische Universität Braunschweig on video-based emotion analysis in intelligent vehicles. I have also interned at vCreaTek, where I worked on the applications of NLP for understanding tabular data.

During my second year, I was involved with some development projects as part of the academic council’s web team: Learner’s Space, iSURP Portal, Scholarship Portal, Creddit/CourseRank Portal (these have either been updated or taken down since then). I also co-developed the Astronomy Club’s website — the existence of which remains publicly unannounced (perhaps they await some celestial alignment). In my free time, I enjoy reading (primarily fiction and history (also fiction?)), writing with unnecessary nested parentheses, traveling, and (occasionally) playing basketball.

The best way to contact me is via email: harshitvarmacse.iitb.ac.in