The Department is looking for a postdoctoral researcher to assist in the analysis of a machine-learning high-resolution global climate emulator that will be developed by a team of climate and computer scientists. The emulator output will be compared to simulations performed with the IFS-NEMO global climate model.
The researcher will also collaborate with the team in the validation of a fine-tuned large-language model (LLM) to provide trustworthy climate information for climate adaptation.
The position is linked to the work performed in the context of the Destination Earth initiative. It involves (1) the contribution to the development and leading the validation of the machine learning-based emulator of a global climate model with the ability to run at eddy-resolving resolutions, (2) comparing the physical performance of the emulator with that of the IFS-NEMO global climate model, and (3) the contribution to the development of the LLM for climate adaptation and leading to its validation.
The candidate will closely collaborate with the teams developing both the emulator and the IFS-NEMO physical model.
No previous knowledge of machine learning techniques is required.
Key Duties
- Lead scientific analyses of the physical performance of the machine-learning emulator, with a special focus on the atmospheric and ocean circulation
- Contribute to performing historical and scenario experiments with the emulator
- Use model evaluation software within the AQUA and/or ESMValTool validation frameworks
- Contribute to the process-based evaluations of the IFS-NEMO climate simulations
- Provide scientific inputs for the development of the eddy-resolving version of the model IFS-NEMO using the evidence from the machine-learning emulator
- Contributing to the fine-tuning of the LLM for climate adaptation and leading its validation
- Preparing and submitting supercomputing access proposals to support the simulations and LLM tuning
- Participate in the BSC contributions to several project deliverables
Data de tancament: Dilluns, 24 Març, 2025
Més posts de Recerca
Cap comentari:
Publica un comentari a l'entrada