Research Fellow in Deep Learning and Large-scale Geophysical Modelling

A three-year postdoctoral research position in machine learning emulators of ice-ocean processes is available at Northumbria University. The position is a part of the project PRECISE: Prediction of Climate Change and Effect of Mitigating Solutions, funded by the Novo Nordisk Foundation. This is an international collaboration project involving several partners including the Niels Bohr Institute as a lead partner. 

You will be developing new approaches based on deep-learning methodologies to create fast emulators of large-scale geophysical models. You will be assisted by several experts in machine learning including Prof Wai Lok Woo at Northumbria University and Dr Hubert Shum at Durham University, and you will be working as a part of an active and large research group at Northumbria University focusing on the Future of Ice on Earth including Profs Hilmar Gudmundsson and Jan De Rydt. 

The specific goal is to develop an emulator of the MIT general ocean circulation model (MITgcm) for a regional configuration of the Southern Ocean to emulate ocean-induced melt rates at the underside of the ice shelves in the Amundsen Sea. Utilising the latest advances in deep-learning, you will create a general framework for emulating large-scale geodynamic simulations. The emulator will be constrained by physical principles and capable of producing output fields near-identical to those of a large-scale ocean circulation model, at a fraction of the computational cost. 

As a part of this large international collaboration, you will be involved in scientific project meetings and will be able to attend conferences and international workshops in glaciology, ice-ocean interactions, and deep learning. The key goal of this post is to develop an ocean dynamics emulator capable of emulating ice shelf melt rates for a range of ocean thermohaline conditions and ice shelf cavity geometries produced by the MITgcm model in a coupled configuration with a dynamical ice-flow model. Prior experience in machine learning is desirable, and a good knowledge of mathematics and physics is required. You will have a PhD in mathematics, physics, or geosciences. 

The application deadline is Sunday 11^th May 2025.

Further information is within the job description. 

We welcome applications from the UK and across the world. Visit our web pages for details about Relocation Assistance.

If you would like an informal discussion about the role, please contact Prof Jan De Rydt and Prof Hilmar Gudmundsson.

Hilmar Gudmundsson (hilmar.gudmundsson@northumbria.ac.uk)

Important dates:

Closing date: 11^th May 2025.

Interview and selection date: w/c 19 May 2025.

Start date: flexible but aiming at September 2025.

Advert information