LEAP’s multi-institutional, trans-disciplinary research community of leading scientists addresses the fast-evolving and complex challenges in climate modeling, climate projection, and climate adaptation. LEAP has a vision to merge physical models with machine learning, and sync the work of climate scientists and data scientists, in order to stimulate convergence between the two disciplines and develop new machine learning algorithms for studying and solving interesting, data-intensive problems. We are also creating a cloud platform, LEAP Pangeo, to provide climate data more widely, and engage with our colleagues in the field and beyond to see what is actually useful.