Four positions for Scientist - Machine Learning, Data specialist and High-performance Computing Engineer roles, to build a foundation model for Earth system science

Your role 

ECMWF is coordinating the WeatherGenerator Horizon Europe project that will build a European generative foundation model of the Earth system. The WeatherGenerator will serve as a new Digital Twin for Destination Earth. It will be based on representation learning and create a general and versatile machine learning tool that models the dynamics of the Earth system by training on a large variety of Earth system data. The WeatherGenerator will be task-independent and will improve results for a wide range of applications when compared to task specific machine learning tools. It will also be more resilient for climate applications, i.e. when the underlying data distributions are changing. it will lead to a significant reduction in computational costs and faster turnaround times compared to existing tools. The WeatherGenerator will be developed in close cooperation with the team working on ECMWF’s machine learning weather forecast model – the AIFS.

To build the WeatherGenerator, we will: 
(1) Collect and use the most important datasets of Earth system science including data from Digital Twins of Destination Earth, selected observations, analysis and reanalysis datasets, and output of conventional Earth system models. 
(2) Build the WeatherGenerator as a novel representation learning-based machine learning tool that trains from petabyte-scale datasets and exploits Europe’s largest supercomputers with training runs with thousands of GPUs.
(3) Engage with the wider weather and climate communities via services and apply the WeatherGenerator for 22 selected applications that can be integrated into the Destination Earth framework. The applications include global and local weather forecasting, local downscaling, data assimilation, model post-processing, and impact applications in the domains of renewable energy, water, health and food. 

The project consortium that will build the WeatherGenerator consists of experts in machine learning, supercomputing and Earth system sciences, and includes industry, small and medium-sized enterprises, and leading operational weather centers. The WeatherGenerator will lead to key innovations in weather and climate science as well as machine learning and enable Europe to establish and defend its leadership in Earth system modelling and associated machine learning tooling. 

Joining the team, you will have the opportunity to create a cutting-edge machine learning model that pushes what is considered possible both scientifically and computationally. The project will also allow you to make a difference in the world by building a practical model that will help solve challenging, global problems in response to the climate crisis. If successful, the WeatherGenerator will likely be continued as a leading European machine learning tool for weather and climate.

We are hiring four experts in machine learning, dataset generation and high-performance computing to realise this ambitious goal and to help revolutionise machine learning applications in weather and climate science. The successful applicants will form the core team at ECMWF to establish the training data, high-performance computing workflows and machine learning design of the atmosphere component of the WeatherGenerator in close collaboration with the partners. WeatherGenerator will work in close collaboration with other machine learning efforts around AIFS and DestinE, generating synergies and share tools and data wherever possible. This phase of the WeatherGenerator project will run from early 2025 to late in 2028 and future phases are foreseen (subject to funding).  

At ECMWF, you will find a passionate community, collectively aiming to build world-leading global Earth system models for numerical weather prediction and climate services.

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 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. 

ECMWF is a global leader in machine learning for Earth system application and investigates machine learning throughout the weather forecast value chain including for observation processing, data assimilation, forecasting and post-processing. ECMWF has also developed a machine learned global forecast model – the Artificial Intelligence Forecasting System (AIFS) – that is used for daily weather predictions.

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.

See   for more info about what we do. 

Your responsibilities 

To complement our existing teams, we are looking for exceptional candidates who fit one or more of the following roles: 

• Data specialist, focused on the collection, curation and generation of large training datasets consisting of a mixture of Earth system data, including reanalysis, model output, observations, and DestinE digital twin data products.
• High-performance computing engineer, focussed on the optimization and scaling of a generative machine learning model based on representation learning for efficient training on hundreds or thousands of GPUs.
• Machine learning specialist, focussed on developing large scale, generative machine learning models that can learn from heterogenous datasets, utilizing state-of-the-art training protocols and architectures.

You can specify in your application for which role(s) you would like to be considered and tailor your application accordingly. We will hire four positions across the three roles and adjust the focus given the strength and potential of individual candidates and the group of applicants.

What we're looking for

Across all roles:

• Dedicated and enthusiastic about teamwork but also self-motivated and able to work with minimal supervision, taking responsibility for a part of the larger 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.  
• Highly organised with the capacity to work on a diverse range of tasks with tight deadlines.
• Significant experience developing in Python or similar languages, and the use of software version control and best practices for software development.

Role 1: Data specialist

• Experience with the generation, handling, and dataset generation of large (TB++) datasets for use in machine learning applications.
• Experience with high-performance computing environments and very large file systems.
• Experience using Earth system data would be an advantage.

Role 2: High-performance computing engineer

• Expertise in at least one deep learning framework (PyTorch, Tensorflow, JAX).
• Experience with machine learning model optimisation to improve memory footprint, training and inference speed.
• Experience with model parallel implementation and optimization of large-scale machine learning models.
• Familiarity with CUDA or Triton would be an advantage.
• A background in Earth system modelling is welcome but not required.

Role 3: Machine learning scientist

• Expertise in at least one deep learning framework (PyTorch, Tensorflow, JAX).
• Experience with generative machine learning, in particular diffusion models, and architectures such as transformers and graph neural networks.
• Familiarity with state-of-the-art in large scale self-supervised learning. 
• Experience with the evaluation of probabilistic 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.

Candidates must be able to work effectively in English and interviews will be conducted in English. A good knowledge of one of the Centre’s other working languages (French or German) would be an advantage.

Education and experience

• Advanced university degree (EQ7 level or above) in a physical, computing, mathematical or environmental science, or equivalent professional experience.
• Experience in the general areas of machine learning and scientific computing.
• Experience in either data processing and efficient data loading, generative machine learning or high-performance computer engineering.
• Experience in Earth system modelling is desirable.

Other information 

Grade remuneration:  The successful candidates will be recruited at Grade A2, according to the scales of the Co-ordinated Organisations and the annual basic salary will be GBP 71,451 (Reading/UK) or EUR 86,824 (Bonn/Germany) NET annual basic salary (ECMWF salaries are exempt of national income tax). In addition to basic salary, ECMWF also offers an attractive package of benefits and entitlements. This position is assigned to the employment category STF-PS  as defined in the ECMWF Staff Regulations. To find out more about working with us and for full details of salary scales and allowances, please visit . 

Starting date:                01 February 2025

Length of contract:     The contract duration is expected to be 42 months, with the possibility of contract extension to 31 January 2029.

Location:                         Reading, UK or Bonn, Germany

Remote work:            As a multi-site organisation, ECMWF has adopted a hybrid organisation model which allows flexibility to staff to mix office working and teleworking. We allow for remote work 10 days/month away from the office, including up to 80 days/year away from the duty station country (within the area of our member states and co-operating states).

Interviews by videoconference (MS Teams) are expected to take place during the second half of October  2024. If you require any special accommodations in order to participate fully in our recruitment process, please let us know. 

To contact the ECMWF Recruitment Team, please email jobs@ecmwf.int.

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. 

ECMWF Member 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. 

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.