Machine Learning Performance Engineer Job Description Your responsibilities What we are looking for Your profile

Machine Learning Performance Engineer | ECMWF - European Centre for Medium-Range Weather Forecasts

Machine Learning Performance Engineer

ECMWF - European Centre for Medium-Range Weather Forecasts Remote | Bonn National A2, EUR 91,754 ( Bonn/Germany) / NET annual basic salary + other benefits level Speaks German, English, French

Application deadline: March 22, 2026

Closed

Apply

Summary by Impactpool

The Machine Learning Performance Engineer will join the HPC Applications team at ECMWF, focusing on benchmarking and optimizing machine learning models for the Destination Earth initiative. The role involves enhancing the performance and scalability of ML models on various EuroHPC systems and contributing to the development of ML workflows. Candidates with experience in optimizing scientific applications on high-performance computing systems are encouraged to apply, even if they lack direct machine learning experience. The position offers a collaborative environment aimed at advancing global weather predictions and climate simulations.

Candidate Requirements:

Advanced university degree in computer science, engineering, or related field

Experience in benchmarking and optimizing large scale scientific/ML software

Knowledge of HPC architectures, especially GPU accelerators

Experience with GPU programming (e.g., CUDA, HIP)

Familiarity with HPC programming models (OpenMP/MPI)

Knowledge of Pytorch and Pytorch Lightning is desirable

Excellent analytical and problem-solving skills

Ability to work autonomously and in distributed teams

Strong interpersonal and communication skills

Fluency in English; knowledge of French or German is a plus

Job reference:

VN26-27

Salary and Grade:

Grade A2, EUR 91,754 ( Bonn/Germany) / NET annual basic salary + other benefits

Deadline for applications:

22/03/2026

Department: Computing

Location: Bonn, Germany

Contract type:

STF-PS

Publication date:

23/02/2026

Contract Duration: Approx. two years to 31 May 2028

Job Description

Your role

We are searching for a highly motivated Machine Learning Performance Engineer to join the HPC Applications team at ECMWF. In this role, you will have a particular focus on benchmarking and optimising machine learning (ML) models within the Destination Earth (DestinE) initiative of the European Commission so that they run efficiently across a variety of EuroHPC systems such as LUMI, Leonardo, MareNostrum 5 and JUPITER as well as ECMWF's own in-house HPC systems. Candidates are encouraged to apply even if they don’t possess any machine learning experience but have experience optimising scientific applications on large scale heterogeneous high-performance computing systems.

At ECMWF, you will find a passionate community, collectively aiming to build world-leading global Earth system models for weather prediction and climate simulations using both physics-based and machine learning approaches. ECMWF have pioneered the AIFS as an operational data-driven forecasting system, and played a key role in the development of Anemoi, an open-source software framework for data-driven earth system modelling. Within DestinE, ECMWF develop earth-system data-driven model components, and oversee the development of a data-driven climate emulator and data-driven regional modelling system, each of which build the Anemoi framework.

The successful applicant’s work with the existing HPC Applications, DestinE and ML teams, will not only contribute to improving the computational performance and scalability of ML models on the world's largest supercomputers but also enable ECMWF to bring and optimise innovative ML approaches into the DestinE Digital Twin Engine workflows, as well as in its operational prediction workflows. This effort supports delivering an AI Earth System Model in DestinE, as well as ECMWF’s strategy of producing cutting‐edge science and world-leading weather predictions and monitoring of the Earth system.

The team

This position is based in the HPC Applications team, responsible for making sure that applications, such as the ECMWF IFS physics-based model and the AIFS machine learning model, run as efficiently as possible on internal and external HPC systems and that they are able to scale across the world's largest supercomputers. The team is also responsible with developing benchmarks for procuring ECMWF's in-house HPC systems used for operational weather forecasting and with providing application support for the operational weather forecasting suites.

About ECMWF

The European Centre for Medium-Range Weather Forecasts (ECMWF) is a world leader in Numerical Weather Predictions providing high-quality data for weather forecasts and environmental monitoring. As an intergovernmental organisation, we collaborate internationally to serve our members and the wider community with global weather predictions, data and training activities that are critical to contribute to safe and thriving societies.

The success of our activities depends on the funding and partnerships of the 35 Member and Co-operating States who provide the support and direction of our work. Our talented staff together with the international scientific community, and our powerful supercomputing capabilities, are the core of a 24/7 research and operational centre with a focus on medium and long-range predictions. We also hold one of the largest meteorological data archives in the world.

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 and the Strengthening Early Earning in Africa (SEWA) 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.

Our vision: The strength of a common goal Our mission: Deliver global numerical weather predictions focusing on the medium-range and monitoring of the Earth system to and with our Member States

ECMWF is a multi-site organisation, with its headquarters in Reading, UK, a data centre in Bologna, Italy, and a large presence in Bonn, Germany, as a central location for our EU-related activities. ECMWF is internationally recognised as the voice of expertise in numerical weather predictions for forecasts and climate science.

About DestinE | Destination Earth

ECMWF is one of the entities entrusted with implementing the European Commission's DestinE initiative , alongside ESA and EUMETSAT, and works closely with over 100 partner institutions across Europe. DestinE delivers high-resolution, configurable digital twins of the Earth system that simulate past, present and plausible future environmental conditions and enable “what-if” scenario exploration. 

ECMWF is responsible for the delivery of these digital twins and of the Digital Twin engine, the software infrastructure needed to power the digital twins on the European HPC Joint Undertaking (EuroHPC) and ECMWF’s supercomputers and to handle and enable access to their data via the DestinE infrastructure, as well as for a range of Artificial Intelligence/ Machine Learning (AI/ML) activities. These include developments towards an ML Earth system model and a range of AI solutions that increase system interactivity and usability of digital twin data.

The third phase of DestinE (June 2026 – June 2028) marks the transition of the Climate Change Adaptation and Weather-induced Extremes Digital Twins, together with the Digital Twin Engine, into sustained operations. During this phase, the focus will be on consolidating, operating, and further evolving these key elements, as well as on maturing the AI Earth-system model components and AI-based solutions developed in phase 2.

The Climate DT is implemented by a partnership led by CSC, currently involving 12 leading climate institutions, supercomputing centres, national meteorological services, academia and industrial partners, through a contract procured by ECMWF.

For more about DestinE and the Climate DT, please see  and .

Your responsibilities

Benchmark ML models on a wide range of EuroHPC systems in the framework of DestinE

Diagnose performance and scalability bottlenecks of DestinE Digital Twin Engine ML-based workflows and come up with solutions to address them

Port, optimise and benchmark DestinE Digital Twin Engine ML workflows on a variety of HPC/ML architectures (e.g., GPUs, TPUs etc)

Support and contribute towards the deployment of ML workflows on EuroHPC systems

Collaboration with colleagues across ECMWF in the ongoing development of DestinE ML data flows required for training and running ML models

Contribute to code review and feature development with stakeholders in DestinE and across ECMWF and Member States

Contribute to wider ML software developments as needed

What we are looking for

Excellent analytical and problem-solving skills with a proactive and constructive approach

Flexibility, with the ability to adapt to changing priorities

Ability to work autonomously and as part of multidisciplinary and geographically distributed teams

Excellent interpersonal and communication skills

Highly organised with the capacity to work on a diverse range of tasks to tight deadlines

Your profile

Education

Advanced university degree (EQ7 level or above) or equivalent professional experience in computer science or engineering, computational science, physics or natural sciences, mathematics, or a related discipline.

Knowledge, Skills and Experience

Experience of benchmarking, optimising, debugging, troubleshooting of large scale scientific/ML software on high-performance computing systems

Knowledge of HPC architectures and in particular, GPU accelerators

Experience with GPU programming at various abstraction levels (e.g., CUDA, HIP, Triton) is highly desirable

Experience with HPC programming models such as OpenMP/MPI and key concepts such as distributed memory and shared memory programming

Knowledge of Pytorch and Pytorch Lightning is highly desirable

We encourage you to apply if you have the relevant profile and motivation to join us, even if you don't precisely meet all these criteria.

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) is an advantage.

Other information

Grade remuneration: The successful candidates will be recruited according to the scales of the Co-ordinated Organisations. In addition to basic salary, ECMWF also offers an attractive benefits package. To find out more about working with us and for full details of salary scales and allowances, please visit .

Starting date:  01 June 2026

Location: Bonn, Germany (Candidates are expected to relocate to the duty station)

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 within a month after the closing date. If you require any special accommodations in order to participate fully in our recruitment process, please let us know.

To

[Veuillez cliquer sur le bouton Postuler ci-dessous pour découvrir comment envoyer votre candidature (E-mail, téléphone, WhatsApp, etc.)]