Postdoc Machine Learning for Data Assimilation in Urban Atmospheric Flow

Job description

The Geoscience and Engineering department of TU Delft is looking for a postdoctoral researcher to contribute to the interdisciplinary project UrbanAIR. The UrbanAIR project is funded by the EU Horizon Europe programme. It involves five departments of TU Delft and 18 international partners across industry, academia and government.

With advanced weather models we can forecast atmospheric circulation at smaller and smaller scales. High-resolution simulations of the urban air flows form the basis of so-called digital twins that support decision making. In digital twins, data assimilation systematically combines numerical models with data to explore different possibilities of what the true system could be like. Machine-learning techniques enable the development of surrogate models that support the data assimilation Given the complex urban geometries and limited observations in cities, a combined machine learning/data-assimilation approach will need to address challenges of data requirements (simulated versus real-world data), training costs, and inference speed.

To tackle these challenges, you will team up with UrbanAIR partners to co-develop and utilise surrogate forward models for prediction of future states or time-step approximations. A promising possibility is to investigate stochastic generative surrogate models to evaluate uncertainty of predictions in addition to the deterministic surrogate models developed by other UrbanAIR partners. By combining surrogate models with existing data-assimilation approaches you and your colleagues will build an efficient tool for Bayesian estimation.

TU Delft is the scientific coordinator of the UrbanAIR project. In this specific position, you will support the scientific coordination. Specifically, you will assist the scientific coordinator in monitoring progress, organising consortium meetings, and reviewing project reporting. You will interact with the project manager at the TU Delft Innovation and Impact Center who carries out the financial and administrative project management. The time allocated for these tasks will be on average four hours per week, with a maximum of eight hours per week.

As a postdoc at TU Delft, you will:

• Develop (stochastic) surrogate forward models for simulation of atmospheric circulation in urban environments
• Co-develop a machine-learning supported data assimilation method based on recently developed methods in machine/deep learning and data assimilation
• Explore innovative strategies to adapt the method to application for urban air circulation
• Implement the resulting algorithm into a digital twin for application in selected cities in Europe
• Support the scientific coordinator in the monitoring of scientific progress, organisation of scientific consortium meetings, and reviewing of scientific content of progress reports.

The four-year EU-funded project UrbainAIR started in January 2025 and will last until December 2028. The postdoc position is available for two years (November 2025-October 2027) fulltime but can also be fulfilled on a part-time basis (0.6 — 0.9 FTE) with a proportionally longer duration of the contract. You will join TU Delft’s Department of Geoscience & Engineering, where you will find a vibrant and collaborative environment. The UrbanAIR project gives you the opportunity to work closely with leading experts and industry partners in machine learning, data assimilation, atmospheric modelling, and numerical simulation. You will be encouraged to learn new skills and develop your own ideas, and you will share your findings at conferences and in peer-reviewed journals. You will receive all the support and coaching you need to develop further your career.

Job requirements 

Our new colleague:

• Holds a PhD degree in data assimilation or uncertainty quantification, preferably in a field related to fluid dynamics and has a strong background in machine learning, especially deep learning, and generative models.
• Has knowledge of fluid dynamics, particularly in turbulent flows (LES and RANS) and experience with numerical simulations and computational methods for partial differential equations.
• Is able to handle and analyze large datasets, both simulated and real-world data and has extensive experience with Python, C++, fortran, or a similar programming language. Working with complex geometries and adapting deep-learning architectures is a plus.
• Is a highly motivated and self-driven researcher, capable of working both independently and as part of an international team. Our new colleague has strong communication skills, also in virtual environments, and an excellent command of written and spoken English

Conditions of employment 

• Duration of contract is 2 years (longer if part-time).
• A job of 24-38 hours per week. 

Salary and benefits are in accordance with the Collective Labour Agreement for Dutch Universities. The TU Delft offers a customisable compensation package, discounts on health insurance, and a monthly work costs contribution. Flexible work schedules can be arranged. 

Will you need to relocate to the Netherlands for this job? TU Delft is committed to make your move as smooth as possible! The HR unit, Coming to Delft Service, offers information on their website to help you prepare your relocation. In addition, Coming to Delft Service organises events to help you settle in the Netherlands, and expand your (social) network in Delft. A Dual Career Programme is available, to support your accompanying partner with their job search in the Netherlands.  . 

Additional information
If you would like more information about this vacancy or the selection procedure, please contact Prof. Femke Vossepoel (F.C.Vossepoel@tudelft.nl).

Application procedure
Are you interested in this vacancy? Please apply no later than 3 July 2025 via the application button and upload the following documents:

• A motivation letter (max. 800 words) outlining how your background fits the job description and requirements and your interest for this particular project.
• A curriculum vitae including, if available, a link to your PhD thesis.
• Academic transcript, including a summary of your PhD thesis.
• Contact information for two referees we can call or email for feedback (no letters are needed at this stage).

After an initial selection, we plan to have a first round of online interviews on July 7th and 8th, 2025. Note that this is very shortly after the closing date of the vacancy. 

We are acutely aware that we are a diverse society and not every talented student will have had the same opportunities to advance their careers. We therefore pledge to fully account for any particular circumstances that the applicants disclose (e.g. parental leave, caring duties, part-time jobs to support studies, disabilities etc.) to ensure an inclusive and fair recruitment process that does not rely purely on common research metrics. Our goal is to help the next generation of scientists thrive in a safe and inclusive environment.

You can address your application to Prof. Femke Vossepoel.

We are looking forward to receiving your application!

Femke Vossepoel (TU Delft)

Anna Gralka (TU Delft)

Please note:

• You can apply online. We will not process applications sent by email and/or post. 
• A pre-employment screening can be part of the selection procedure.
• For the final candidates, a knowledge security check will be part of the application procedure. For more information on this check, please consult Chapter 8 of the National Knowledge Security Guidelines. We carry out this check on the basis of legitimate interest.
• Please do not contact us for unsolicited services.