Sattwik Kumar Sahu

I develop large-scale multimodal models for zero-shot robot navigation and generative control, deploying them on real hardware, with a focus on self-supervised representation learning and latent predictive world models for robotics.