Scientist for Stochastic Parametrisation and Differentiable Physical Processes

Your role

We are looking for a highly motivated (Senior) Scientist to work on the representation of uncertainty in ECMWF’s ensemble forecasts of the Integrated Forecasting System (IFS) and maintain the tangent linear (TL) and adjoint (AD) code for the IFS physical parametrisations used during the minimisation process of 4DVar data assimilation. This includes current operational stochastically perturbed parameterisations (SPP) scheme, use of singular vectors, uptake of initial conditions from the ensemble data assimilation system, and support for the Artificial Intelligence Forecasting System (AIFS).

Responsibilities

• Enhance representations of uncertainties (e.g., SPP stochastic parametrisation) across forecast lead times and km‑scale model simulations.
• Maintain and update TL/AD model code for physical parametrisation schemes of the IFS, including exploration of new methods such as automatic differentiation and deep learning emulation.
• Support developments of the AIFS ensemble system by providing insights into the representation of physical processes and generating training datasets for data‑driven ensemble models.

What We're Looking For (Qualifications)

• Excellent analytical, problem‑solving, and proactive approach to model and tool improvement.
• Excellent interpersonal and communication skills.
• Self‑motivated with ability to work independently yet collaboratively.
• Strong documentation and communication of scientific results.
• Highly organised and able to manage diverse tasks under tight deadlines.

Your Profile - Education, Experience, Knowledge And Skills

• Advanced university degree (EQ7 level or above) in a physical, mathematical or environmental science, or equivalent professional experience.
• Experience in Earth system modelling, contributions to code, large‑scale simulations on modern supercomputing environments.
• Experience in stochastic parameterisation schemes and/or generation of tangent linear and adjoint model code is desirable.
• Expertise in atmospheric physical processes, numerical weather prediction and operational weather prediction methodology is desirable.
• Ability to work effectively in English.

Benefits

• Position may be recruited at A2 or A3 grade, according to ECMWF staff regulations.
• Flexible teleworking policy; hybrid working model with up to 80 days/year remote work.
• Support for relocation to duty station (Reading, UK or Bonn, Germany).
• Contract for four years with possibility of extension.
• Salary scales available on ECMWF website; benefits package includes generous allowances.

Location

Reading, UK or Bonn, Germany; relocation support provided.

Equal Opportunity Statement

ECMWF is committed to an inclusive environment and provides equal opportunities for all applicants.