Machine Learning Scientists and Machine Learning Engineers (3 positions)

We are looking for three highly motivated Machine Learning Scientists and Engineers to develop world‑leading machine‑learning models for weather and climate and efficient, scalable and sustainable software infrastructure ready for the Exascale Era. The work requires both technical expertise to create state‑of‑the‑art and sustainable models and excellent soft‑skills to collaborate seamlessly in an international, interdisciplinary environment.

Roles:

• Role A: Machine Learning Scientist (Modelling team) - Design, implement, train and scale machine‑learning models for long‑time integrations, with a focus on advanced time‑stepping techniques and physical consistency. Develop models for the sub‑seasonal and seasonal Earth system components (e.g., ocean, sea‑ice, land surface, hydrology). Investigate model parallelisation approaches and HPC efficiency on EuroHPC systems. Experience with model design, evaluation and by extension > months to years time‑scales is advantageous.
• Role B: Machine Learning Engineer (Engineering team) - Design, build and maintain reproducible machine‑learning pipelines and workflow orchestration for Anemoi and AI Factories. Manage dependencies, orchestrate complex multi‑step workflows and work with HPC or large‑scale GPU environments. Familiarity with job schedulers such as SLURM, distributed training frameworks (e.g., PyTorch DDP) and CI/CD or orchestration tools (Airflow, Prefect) is a plus.
• Role C: Machine Learning Scientist (Modelling team) - Design and evaluate machine‑learning models for a machine‑learned Earth system model, focusing on training and inference performance. Scale machine‑learning tools to large HPCs with hundreds or thousands of nodes. Apply machine‑learning approaches to Earth system modelling challenges.

Across all roles:

• Excellent analytical, problem‑solving and communication skills.
• Self‑motivated, able to work with minimal supervision, and collaborative.
• Strong documentation and effective communication of scientific results.
• Proficiency with standard software development tools (git, coding best practices, tests, code review).
• Highly organised and capable of handling diverse tasks to tight deadlines.

Your profile:

• Advanced university degree (EQ7 level or above) in physical, mathematical, data/machine‑learning, or environmental science, or equivalent experience.
• Experience in machine‑learning or machine‑learning engineering, including best practices for software development.
• Experience in Earth system modelling is desirable but not mandatory.
• Experience in HPC or large‑data science projects is desirable.
• Must be able to work effectively in English.

Remuneration and Start Date:

Grade remuneration will follow the ECMWF scales of the Co‑ordinated Organisations. Salary scales and allowances details are available at the ECMWF website. Starting date is as soon as possible.

Location:

Successful candidates will relocate to Bonn, Germany, or Reading, UK. ECMWF operates a hybrid model allowing office or teleworking flexibility.

Equal Opportunities:

ECMWF is committed to an inclusive workplace that embraces diversity and provides equal opportunities for all, without distinction by race, gender, age, marital status, social status, disability, sexual orientation, religion, personality, ethnicity or culture. We welcome applications from nationals of ECMWF Member States, Co‑operating States, the European Union and Ukrainian nationals. In exceptional cases, applications from other countries may be considered.