Himanshu Sahni
DeepMind
About
Himanshu Sahni is a Research Scientist at DeepMind working at the intersection of reinforcement learning and generative modeling. His research spans scalable RL algorithms, reward learning, and efficient diffusion-based language modeling. He completed his PhD in Artificial Intelligence at Georgia Tech, where he focused on sample-efficient reinforcement learning. Himanshu was a core contributor to Gemini Diffusion, presented at Google I/O 2025, and has conducted research across industry labs including Tesla, Facebook, and Microsoft. He is passionate about bridging algorithmic insights with real-world deployment.