Gagan Jain

About

I’m an Applied Scientist at Microsoft AI. My primary research interests are in generative models and the foundations of deep learning. I also maintain a keen interest in diffusion models and unsupervised post-training for large language models, particularly via reinforcement learning.

Previously, I was a pre-doctoral researcher in the Machine Learning and Optimization team at Google DeepMind, Bangalore, India, advised by Dr. Sujoy Paul, Dr. Prateek Jain, Dr. Arun Suggala, and Dr. Karthikeyan Shanmugam, working on efficient machine learning models, reinforcement learning for diffusion models, and theoretical guarantees for collaborative bandit approaches.

Even earlier, I graduated with a B.Tech (Hons.) in Mechanical Engineering from the Indian Institute of Technology Bombay, with minors in Computer Science and Machine Intelligence. During my undergrad, I focused on robotics (ground and aerial vehicles), especially autonomous driving.