Welcome!

Hello, I’m Sarvesh Gharat, a Ph.D. student at the Centre for Machine Intelligence and Data Science (CMInDS), Indian Institute of Technology Bombay (IIT Bombay). I completed my Bachelor’s degree in Electronics and Telecommunication Engineering at VIIT Pune.

My research lies at the intersection of online learning and generative AI. In online learning, I focus on multi-armed bandits and PAC decision-making, studying how to make reliable decisions under uncertainty. In generative AI, I apply these ideas to large language models (LLMs), developing methods for efficient model selection and improved reasoning through adaptive prompting and inference-time strategies.

More recently, I have become interested in LLM alignment, multi-agent systems, and AI Scientists. I am particularly interested in building reliable agents that can interact and collaborate effectively, as well as systems that can automate parts of the scientific process—from generating hypotheses and designing experiments to evaluating results and iteratively improving AI methods.

I have been fortunate to work with inspiring mentors and collaborators, including through internships at Google DeepMind and Adobe Research, which have shaped my perspective on both theoretical and applied AI research.

Outside of academia, I enjoy probability theory for the elegance it brings to understanding uncertainty. I’m also passionate about sports — especially football and cricket — which I value for the teamwork, strategy, and competitive spirit they foster.

I’m grateful for the opportunities and guidance that have shaped my journey so far, and I’m excited to continue exploring new directions in AI research.

Thank you for visiting my corner of the web — feel free to connect if you share similar interests!