Postdoctoral position in simulation-based inference for particle physics

Do you want to work with machine learning and simulation-based inference for the search for dark matter (or other "invisible" signals of new physics) at Large Hadron Collider, in an international environment with competent and pleasant colleagues? Welcome to apply for a postdoctoral position at Uppsala University.

This is an abbreviated version of the ad. The full advertisement is available on Uppsala University's website, uu.se/jobb.

The Department of Information Technology is Uppsala University's third largest institution with just over 350 employees. The position is located at the Department of Computational Science (TDB), one of the world's largest research environments in computational science with extensive activities in areas such as machine learning, optimization and high performance calculations, and an important part of eSSENCE and SciLifeLab.

The person employed is part of the research group Scientific Machine Learning at TDB and SciLifeLab (associate professor Prashant Singh), who develops methods and software for simulation-based inference, generic models and robust machine learning, in close collaboration with the group of theoretical particle physics at the Department of Physics and Astronomy (professor Stefano Moretti), who researches phenomenology beyond the standard model and dark matter and is a member of the CMS experiment at CERN. The project establishes a new cross-faculty collaboration where new methods for simulation-based inference are developed and applied directly in realistic analyses of colliding data. The postdoc is supervised jointly by both groups and receives support for conference trips as well as access to national HPC resources (NAISS) and local GPU infrastructure.

The position is part of the eSSENCE research school in data-intensive science, an arena where experts in computational science, computer science and computer technology work closely with researchers in computer-driven sciences, industry and society. eSSENCE is a strategic research collaboration within e-science between Uppsala University, Lund University and Umeå University.

Project descriptionThe search for dark matter at LHC means that high-dimensional collision data is compared to detailed simulations whose limelihood cannot be calculated, only sampled. Simulation-based inference (SBI) attacks this by training neural networks, such as generic models based on flow matching, on simulated events. The project aims to develop efficient, robust and calibrated SBI methods that take into account event selection and systematic uncertainties, and to demonstrate them on realistic searches on a large scale. The project is primarily based on simulated data, but it is also possible to work with open data from ATLAS and/or CMS. The methods are general and applicable far beyond particle physics.

Tasks Research within the project, including method development, implementation, large-scale computational experiments and publishing, as well as presentation at international conferences, contributions to the group's open software, participation in eSSENCE research school activities and participation in student supervision. A limited percentage of teaching can be included (maximum 20 %).

Qualification requirements Doctoral degree in Machine Learning, Computational Science, Statistics, Physics or a related field, or a foreign degree that is considered equivalent to a doctoral degree in these areas. The degree must be completed no later than the time when the employment decision is taken. In particular, the person who has completed a degree not exceeding three years ago should be considered. When calculating the three-year framework period, the starting point is the closing date for applications. If there are special reasons, such a degree may have been completed earlier. Special reasons include leave due to illness, parental leave, positions of trust in trade unions, etc.

Documented experience in machine learning, especially deep generic models and/or probabilistic modelling, as well as very good programming skills in Python and a modern framework for deep learning (e.g. PyTorch or JAX) are required. Good knowledge of English in speech and writing is required. The candidate shall clearly document a high degree of self-motivation in the application. Great emphasis is placed on personal qualities such as creativity, accuracy, a structured way of working and the ability to work independently and in an interdisciplinary team.

Wishful/meriting otherwise Experience of simulation-based inference, normalizing flows, flow matching or diffusion models; of particle physics (e.g. MadGraph, Pythia, Delphes or analysis of LHC data); of large-scale training on GPU/HPC systems, active learning and open software development are meritorious, as are publications at leading machine learning conferences or physics journals. Teaching experience (e.g. teaching, supervision, mentoring or other educational activities) is meriting but not a requirement; particular emphasis is placed on activities that support students' learning in computer science, information technology or related subjects.

Application The application must include: CV; copy of relevant grade documents (in Swedish or English); publication list; up to five selected publications in electronic format; a research description of previous and current research (max 1 page) and a proposal for future activities (max 1 page); contact information for two references.

In this recruitment we have replaced the personal letter with questions that you answer in connection with your application. Responses are used as part of the selection process.

If the employment is limited in time for two years according to a central collective agreement. The scope is full-time. Admission 1 November 2026 or by agreement. Location: Uppsala.

Information about the position is provided by: Senior Lecturer Prashant Singh, prashant.singh@scilifelab.uu.se; Professor Stefano Moretti, stefano.moretti@physics.uu.se; Head of Department Elisabeth Larsson, elisabeth.larsson@it.uu.se.

Welcome with your application no later than Thursday 15 October 2026, UFVPA 2026/2734.