Yash Thube
I care about making machines intuitively understand space, time, physics and adapt to our physical world. Broadly, my interests revolve around 3D Computer Vision, Reinforcement Learning, and AI for science.Works shaping my thinking: JEPA/V-JEPA (predictive architectures), NVIDIA Cosmos , Physical Intelligence (world foundation models), DeepMind’s Dreamer , Genie series (imagination-based), Aplhafold & Alphagenome (biology), Sakana AI’s work in open-endedness, World Labs, Runway (spatial intelligence) and many more!
Currently working on vision at Future Sportler, a sports tech startup based in India. Always open to collaborations!