Postdoctoral Researcher In ai For Ocean-Atmosphere Interactions During Tropical Cyclones And High Latitude Storms M - F H/F Ifremer

Closure date for receipt of applications : 15/11/2025

Who are we?

You will be hosted at Ifremer within Laboratory of Satellite and Interface Air-Sea. Ifremer provides an ideal environment for the development of state-of-the-art deep learning analysis for ocean-atmosphere processes, with access to the Datarmor HPC system equipped with high-end GPUs and support from machine learning engineers.

The position is part of a large-scale effort at Ifremer to develop learning-based solutions for ocean digital twins and contributes directly to the AI action of the national PPR Ocean & Climate. You will benefit from a dynamic collaborative environment.

The supervising team includes ocean modelling and AI researchers and TCs' observation specialist from the IREMER and the SHOM.

What will your mission BE?

This postdoctoral project focuses on developing novel AI-driven generative models to better understand and predict the ocean's response and feedback to tropical cyclones (TCs) and high latitude storms. The goal is to leverage cutting-edge machine learning techniques to explore how the coupled ocean-atmosphere system, particularly the upper ocean's response, and how it modulates the dynamics of TCs and high latitude storms.

What will your activities BE?

- Utilize nextGEMS high-resolution climate simulations as the primary data source for training advanced AI models.
- Develop and train neural field representations to create compact and continuous representations of TC states and their surrounding oceanic environment.
- Evaluate the transferability of the developed model to high latitude storms.
- Train probabilistic diffusion models in the learned latent space to generate realistic ensembles of TC and storms evolutions.
- Design and utilize an Observing System Simulation Experiment (OSSE) framework to test the AI models and explore optimal observing strategies.
- Assess the realism of nextGEMS simulations against curated ancillary real-world observational datasets (e.g., satellite data from MAXSS/CyclObs/SWOT, Argo floats).
- Investigate the transferability of models trained on simulations to real observations, contributing to the PPR data challenge.
- Publish research findings in peer-reviewed journals and present at international conferences.

Who do you work with?

Internal collaborative relationship

- You will work closely with the recently formed digital twin unit interacting with both Ocean Research and Scientific computing teams.
- The position allows joining the MAXSS project funded by the European Space Agency.
- You will work within the Ifremer/LOPS/SIAM department and the Inria Odyssey team (IMT-Atlantique, IINRIA), collaborating with researchers, engineers, and students.

External collaborative relationship

- You will be co-supervised by modelling and AI researchers from the SHOM who will bring valuable expertise on the data and methods
- The project involves collaboration with partners in long-term projects such as ESA MAXSS, CyclObs, and sentinel MPC projects, Earth Explorer 10 Harmony preparation, and the Copernicus Imaging Microwave Radiometer (CIMR) preparation.

How are your activities organised?

- Full-time job.
- The position is based in Brest, France.
- The role includes a budget for publications and to attend international symposiums.
- 24-month postdoctoral position