Research Software Engineer – User Configurable AI‑Ready Datasets

About the role

ECMWF is seeking a Research Software Engineer to develop data‑infrastructure services that enable AI‑ready meteorological datasets for the EU Destination Earth (DestinE) initiative. You will turn the Zarr‑FDB prototype into a production‑grade, user‑facing service that supports machine‑learning model training and scientific exploration.

Key responsibilities

• Design, implement and optimise tools and workflows for assembling large‑scale datasets for ML training.
• Extend the Zarr‑FDB prototype into a robust, observable, and high‑performance service.
• Collaborate with production teams, service owners, infrastructure providers and researchers across multiple data centres.
• Ensure scalability and reliability of data services in HPC and cloud environments.
• Contribute to the Data Management Services Team’s semantic data‑handling strategy.

Required profile

• Research Software Engineer (A2) with experience in scientific data processing and ML pipelines.
• Strong problem‑solving skills and ability to innovate in a collaborative, multi‑disciplinary environment.
• Familiarity with large‑scale meteorological datasets and digital‑twin concepts.

Required skills

• Python programming
• Zarr data format
• ECMWF Field DataBase (FDB)
• Polytope and Gribjump tools
• High‑performance computing (HPC) environments
• Cloud platforms and containerisation
• Version control (Git) and CI/CD practices

What we offer

• Opportunity to work on cutting‑edge climate‑science data services.
• Collaboration with leading researchers and engineers in meteorology.
• Access to state‑of‑the‑art HPC and cloud infrastructure.