interesting research ideas. Experience training large-scale foundation models (VLMs, text-to-video models, etc) is desirable... both fundamental research advances and practical capabilities. Our work spans computer vision, multi-modal learning, and robotic...