The Climate Physics group at ENS de Lyon, France, is inviting applications for a postdoctoral position (up to 3 years, ERC IceAblation) on Machine learning emulators of polar ocean dynamics.
Can neural networks emulate chaotic, multi-scale ocean flows with tipping points beneath Antarctic ice shelves, generalize across regimes, and stay stable when coupled online to a numerical solver? We have ready-to-use NEMO simulation data and offer freedom to explore physics-constrained learning, diffusion/flow matching models, foundation models, and online training.
Prior experience in ML/AI is needed but we do not require a background in oceanography. Review of applications starts November 2nd, 2026. Details available here.
