The role

The position will develop cutting-edge machine learning applications to support future climate projections. This will allow you to apply and extend the latest generative machine learning approaches and train these at scale on very large datasets. Simultaneously, you will help to define the state-of-the-art in Earth system modelling and support Europe’s efforts to better understand and adapt to climate change.
Machine learning techniques for weather forecasting have made tremendous progress in the last two years, with state-of-the-art models now providing skill comparable to the best equation-based models at a fraction of the compute costs. For climate, large-scale machine learning techniques are still in their infancy and have so far mainly played a supporting role. The open position gives you the possibility to contribute to the development of large generative machine learning for climate applications. The positions is funded by two related projects, Warm World ICON-Rep and EXPECT.
The Warm World ICON-Rep project will extend the Atmo Rep model (https://github.com/clessig/atmorep), a large-scale, self-supervised representation learning model for atmospheric dynamics. In particular, you will enable native support for ICON mesh data in Atmo Rep and than train the adapted model on ICON climate simulations using large scale compute infrastructure, including JUPITER - the first European exa-scale supercomputer at the Jülich Supercomputing Center. With the trained model, it will be explored to what extent an interpolation or extrapolation of climate scenarios is possible, e.g. if training on a small number of CMIP scenarios will allow to generate samples also from other ones not seen during training. Such a capability would hold enormous potential to speed-up and improve climate projections.
The EXPECT project develops European climate change projection and attribution capabilities. The position will support this by implementing a novel downscaling model that can take low-resolution climate model simulation data as input and produce high-resolution data that provides climate information at the local scale where most adaptation and mitigation efforts take place. The localized data should be statistically consistent with high-resolution climate simulation, e.g. from the EERIE project (https://eerie-project.eu), the Destination Earth Climate Digital Two and Next GEMS (https://nextgems-h2020.eu) but be produced at a small fraction of the computational costs.
The position gives you the possibility to be part of and shape the exciting developments on machine learning for Earth System Modelling that are currently taking place at ECMWF (e.g. https://www.ecmwf.int/en/about/media-centre/aifs-blog), in the Warm World and EXPECT projects, and the wider community.

The team

The position will be part of the Earth System Modelling section at ECMWF and tightly linked to the different machine learning efforts at the Centre, e.g. the AIFS, ECMWF's data-driven forecasting model and the Weather Generator project.

About ECMWF

The European Centre for Medium-Range Weather Forecasts (ECMWF) is a world-leader in weather and environmental forecasting. As an international organisation we serve our members and the wider community with global weather predictions and data that is critical for understanding and solving the climate crisis. We function as a 24/7 research and operational centre with a focus on medium and long-range predictions, holding one of the largest meteorological data archives in the world. The success of our activities builds on the talent of our scientists and experts, strong partnerships with 35 Member and Co-operating States and the international community, some of the most powerful supercomputers in the world, and the use of innovative technologies and machine learning across our operations.
ECMWF has also developed a strong partnership with the European Union and has been entrusted with the implementation and operation of the Destination Earth Initiative and the Climate Change and Atmosphere Monitoring Services of the Copernicus Programme. Other areas of work include High Performance Computing and the development of digital tools that enable ECMWF to extend provision of data and products covering weather, climate, air quality, fire and flood prediction and monitoring.
ECMWF is a multi-site organisation, with a main office in Reading, UK, a data centre/ supercomputer in Bologna, Italy, and a large presence in Bonn, Germany. We appreciate the need for flexibility in the way our staff work. We have adopted a hybrid work model that allows flexibility to staff to mix office working and teleworking, including away from the duty station for up to 10 days/month (within the area of our member states and co-operating states).
Seewww.ecmwf.int/ for more info about what we do.

About the projects

The Warm World project (https://www.warmworld.de) develops the next generation of climate models and the required computational infrastructure for this, funded by the German federal ministry for education and research. The ICON-Rep project, which is part of the Warm World Smarter stream, adapts the existing Atmo Rep model for Warm World data and applications. ICON-Rep-Data focuses on the model development and climate scenarios. The sister project ICON-Rep-Data targets the compression of high-resolution data using Atmo Rep, with both models being linked through the use of Atmo Rep.
The EXPECT project develops a prototype operational capability for integrated attribution and prediction of climate phenomena by exploiting novel data and technologies to provide trustworthy assessments and predictions of regional climate change including extremes, funded by the European Union through a Horizon project. The open position supports the project through a machine learning-based downscaling application.

Main duties and key responsibilities

ICON-Rep:

• Adapt the existing Atmo Rep model to support the native ICON-grid and explore the possibility to do climate scenario interpolation and extrapolation
• Support overall Atmo Rep model developments, including for in the Warm World sister project ICON-Rep-Model at the Jülich Supercomputing Centre

EXPECT:

• Implement downscaling of climate data using a generative machine learning model using high-resolution model output from Destin E, EERIE, and next GEMS
• Support coordination of EXPECT’s theme one data for new climate knowledge

What we are looking for

• Dedicated and enthusiastic about teamwork but also self-motivated and able to work with minimal supervision, taking responsibility for the project
• Excellent analytic skills to analyse problems and methodologically develop potential solutions and empirically evaluate them
• Excellent interpersonal and communication skills, and ability for efficient documentation and communication of scientific results
• Significant experience developing in Python or similar languages, and the use of software version control and best practices for software development
• Substantial experience with at least one deep learning framework (usually either Py Torch or JAX), in particular development and tuning of new architectures and training protocols
• Experience with the evaluation of machine learning applications and running of large ablation studies to determine optimal architecture hyperparameters
• A background in Earth system modelling is welcome but not required. Experience with large HPC systems is an asset

Education

• Advanced university degree (EQ7 level or above) in a computing, physical, mathematical or environmental science, or equivalent professional experience

Experience, Knowledge and Skills

• Experience in the general areas of machine learning and scientific computing
• Experience in machine learning and particular in machine learning model development
• Ability to deliver ready-to-use code
• Experience with generative machine learning and Earth system modelling is desirable
• Candidates must also be able to work effectively in English . A good knowledge of one of the Centre’s other working languages (French or German) is an advantage

We encourage you to apply even if you don’t feel you meet precisely all these criteria.

Other information

Grade remuneration: The successful candidates will be recruited at the A2 grade, according to the scales of the Co-ordinated Organisations. The position is assigned to the employment category STF-PL as defined in the ECMWF Staff Regulations. Full details of salary scales and allowances available on the ECMWF website at www.ecmwf.int/en/about/jobs.
Starting date: as soon as possible
Candidates are expected to relocate to the duty station.
Interviews will be conducted by videoconference (MS Teams).
Successful applicants and members of their family forming part of their households will be exempt from immigration restrictions.

Who can apply

Applicants are invited to complete the online application form by clicking on the apply button below.
At ECMWF, we consider an inclusive environment as key for our success. We are dedicated to ensuring a workplace that embraces diversity and provides equal opportunities for all, without distinction as to race, gender, age, marital status, social status, disability, sexual orientation, religion, personality, ethnicity and culture. We value the benefits derived from a diverse workforce and are committed to having staff that reflect the diversity of the countries that are part of our community, in an environment that nurtures equality and inclusion.
Applications are invited from nationals from ECMWF Member States and Co-operating States, as well as from all EU Member States. In these exceptional times, we also welcome applications from Ukrainian nationals for this vacancy. Applications from nationals from other countries may be considered in exceptional cases.
ECMWF Member States and Co-operating States are: Austria, Belgium, Bulgaria, Croatia, Czech Republic, Denmark, Estonia, Finland, France, Georgia, Germany, Greece, Hungary, Iceland, Ireland, Israel, Italy, Latvia, Lithuania, Luxembourg, Montenegro, Morocco, the Netherlands, Norway, North Macedonia, Portugal, Romania, Serbia, Slovakia, Slovenia, Spain, Sweden, Switzerland, Türkiye and the United Kingdom.