Machine Learning Scientist for Weather Forecasting

The Danish Meteorological Institute (DMI) is seeking a talented machine learning scientist with a strong interest in predicting weather to support the green transition through the development and application of innovative data-driven models. This full-time position is within the field of environmental sciences, with a special focus on meteorology and Numerical Weather Prediction (NWP).

International Collaborations DMI actively participates in several exciting international initiatives in the field of limited-area NWP, and is a leading contributor in ongoing research to build kilometre-resolution machine learning forecasting models. This involvement encompasses the deployment and development of advanced weather models, featuring high-resolution simulations and data assimilation. As a member of our team, you will collaborate with international research partners in EUMETNET/ACCORD/HIRLAM/United Weather Center (UWC) to develop data-driven forecasting models, in particular anemoi (https://github.com/ecmwf/anemoi-models) and neural-lam (https://github.com/mllam/neural-lam). Your responsibilities will include contributing to a team in the Weather Research Department, with a focus on model development, in addition to the core modeling work.

Qualifcations and experience

• Educational Background: A Master's or Ph.D. degree in applied mathematics, atmospheric sciences, applied physics, data science, or a closely related field is required. Your academic background should reflect a solid foundation in quantitative and computational disciplines.
• Machine Learning Expertise: Substantial experience and a proven track record in scientific machine learning is a necessity, and particularly experience in the context of environmental sciences will be a benefit. Experience with applying state-of-the-art techniques on both structured (e.g. gridded image data) and heterogeneous (e.g. point-wise observations) to produce predictions, particularly probabilistic ones, will be a benefit. In particular, we envision making heavy use of graph-based learning, Bayesian inference techniques, generative learning using diffusion models as well as contemporary gridded techniques such as UNets, both in supervised and self-supervised settings.
• Scientific Programming: Proficiency in scientific programming is essential. You should have demonstrable skills in software development for data analysis and modelling. Strong proficiency in scientific analysis and modelling physical systems in Python is essential, familiarity with Julia and Fortran is highly advantageous. Experience with continuous integration and deployment (CI/CD) practices, tools, and methodologies is an advantage, although not a strict requirement.
• Machine Learning: In-depth experience with frameworks such as pytorch, pytorch-lightning, jax, Flux.JL, Zygote.JL and MLFlow will be a benefit.
• Data Handling: Strong data management and analysis skills, including data pre-processing and post-processing, are vital for success in this role.
• Linux and Scripting: Proficiency in Linux operating systems and scripting is necessary for working in our research environment.
• Language Skills: Proficiency in both oral and written English is essential to collaborate effectively with our international partners and convey your research findings. As our new colleague you will be encouraged to learn Danish, although it is not a requirement.
• Innovative Thinking: We seek individuals who are naturally curious, creative, and highly motivated. Your ability to think outside the box and propose innovative solutions will be a valuable asset to our team.
• Collaborative Mindset: Effective communication and strong teamwork skills are fundamental. You will be working closely with colleagues, both within DMI and with international research partners, to advance our understanding of weather prediction. Additionally, a keen ability to develop, implement, and clearly communicate structured working processes is essential to ensure smooth collaboration and shared understanding across teams.

About the department The Weather Research Department at DMI is comprised of approximately 60 scientists, with 25 of them working in the weather modelling section. We boast a diverse and internationally recognized staff, reflecting the collaborative nature of our research. DMI offers a flexible working week of 37 hours, inclusive of a paid lunch break, and provides 6 weeks of paid vacation annually.

If you are passionate about the intersection of machine learning, meteorology, and numerical weather prediction and are eager to contribute to the development of new data-driven models for accurate and near real-time prediction of weather, we encourage you to apply for this exciting opportunity at DMI. Your work will have a direct impact on enhancing weather forecasting capabilities, supporting the green transition and ensuring the safety and well-being of communities in Denmark and Greenland. Join us in pushing the boundaries of weather prediction and environmental science.

About DMI DMI is the national weather forecast and warning authority for Denmark and Greenland. With a rich tradition of operating and advancing models for numerical weather prediction, DMI stands at the forefront of high-resolution numerical weather forecasting. The forecasts generated by our operational numerical models are of utmost importance in safeguarding Danish society against weather-related risks. This responsibility also extends to providing input for oceanographic models and a forthcoming hydrological warning service, which predicts critical water levels and potential flooding in Denmark. Additionally, the growing emphasis on sustainable energy sources, such as wind, water, and solar power, is heightening the demand for detailed reanalysis and forecasts.

We have approximately 380 employees, most of whom work at DMI’s new headquarters at Sankt Kjelds Gård in Østerbro. The new place have interior design that allows for an activity-based way of working.

Employment Employment and salary will be according to the Danish Law and Agreements. It will be possible to apply for an addition to the basic salary. Researchers who are recruited from abroad, and who are able to meet certain conditions, have the opportunity to work in Denmark for a period of 60 months under a special 26 % tax rate without deduction, instead of working under the regular income taxation. The positions are limited for a duration of 3 years, with the possibility for extension. Permanent positions are possible from the start of contract if skills and experiences advocate for it.

For further details about the position, contact the Head of Weather Modeling Rune Carbuhn Andersen, telephone +4530229985 or by e-mail [email protected] or contact ML expert Leif Denby, telephone +4523724886 or by e-mail [email protected].

At DMI, diversity is an important value for us, because we believe that an inclusive and versatile work environment strengthens task fulfilment. We work actively with diversity in our employee composition, which is reflected in our inclusive workplace with a balance between work life and family life. We encourage everyone to apply for the position regardless of age, gender, sexuality, religion or ethnicity.

Application The application must include a motivated cover letter, a curriculum vitae, relevant exam papers and diplomas and contact for relevant professional references. The application must be received at DMI at the latest 15. September 2024, 12:00 CET..

We expect to carry out interviews on the 20. and 23. September. A second interview may be arranged on the 26. or 27. September. DMI may choose to conduct interviews online.

To apply for the position, please use the below link "Apply for Position", and follow the instructions.