No summary provided for this position.

This fully funded postdoctoral position at KU Leuven focuses on developing remote sensing driven downscaling solutions for Antarctica, leveraging deep learning and climate science to improve climate projections in polar regions. The successful candidate will pioneer physically informed super-resolution models to enhance the spatial resolution and physical consistency of key climate variables, such as Surface Mass Balance (SMB) and surface melt, which are crucial for understanding ice sheet stability and sea level rise. The project integrates multi-source satellite data and geophysical constraints (elevation, albedo, etc.) into advanced climate downscaling frameworks, utilizing high-performance computing resources and collaborating with the Earthmapps research team. Responsibilities include developing and refining deep learning architectures, implementing physical constraints using both soft and hard techniques, validating models with in-situ and satellite-derived data, and applying downscaling methods to regional climate model outputs. The postdoc will also contribute to open-source software development, publish results in high-impact journals, and participate in international collaborations and research visits. Supervision is provided by Prof. Stef Lhermitte, and the position offers opportunities to help supervise related MSc and PhD projects, contribute to teaching activities, and engage with impactful climate science initiatives relevant to the IPCC and Digital Twin projects. Applicants should have a PhD in a relevant field, strong programming and analytical skills, experience with deep learning frameworks, and an interest in climate applications of AI. KU Leuven provides a supportive, inclusive research environment and encourages diversity. The application deadline is November 4, 2025.