PhD Student in Machine Learning with a Focus on Uncertainty Quantification in Large Language Models for Healthcare Research Uppsala University in Sweden

Uppsala University invites applications for a funded PhD position in machine learning, focusing on uncertainty quantification in large language models for healthcare research. The position is based at the Department of Information Technology, a leading research and education hub with a vibrant international environment. The Division of Systems and Control, where the project is hosted, specializes in developing theory and tools for systems that learn, reason, and act using data, mathematical models, and algorithms. Research at the division spans machine learning, optimization, control theory, and applied mathematics, with applications in medicine, energy, biomedical systems, neuroscience, and safety.

This PhD position is part of the prestigious DDLS Fellows program, a national initiative funded by the Knut and Alice Wallenberg Foundation to advance data-driven life science in Sweden. The DDLS program aims to recruit and train the next generation of computational and data scientists, with a strong focus on interdisciplinary research and international collaboration. The program covers four strategic areas: cell and molecular biology, evolution and biodiversity, precision medicine and diagnostics, and epidemiology and biology of infection. The research school will expand in 2026, recruiting new academic and industrial PhD students.

The project centers on developing uncertainty-aware methods for large language models (LLMs) in healthcare, specifically targeting the extraction and quantification of clinical information from unstructured medical text. Current LLM approaches often lack principled uncertainty quantification, which limits their reliability in medical applications. The research will focus on quantifying and evaluating uncertainty in both numerical and categorical predictions from medical reports, integrating these uncertainties into downstream probabilistic time-to-event models. Applications will emphasize prostate cancer, leveraging large-scale clinical registry data and unstructured medical text. The project is highly interdisciplinary, involving close collaboration with clinicians and epidemiologists.

As a doctoral student, you will primarily devote your time to graduate education and research, with the possibility of up to 20% departmental duties such as teaching and administration. The position offers a unique opportunity to gain international experience, supported by a wide network of collaborators at institutions like the University of Oxford, University of Melbourne, and UCLA.

Applicants must hold a Master’s degree in applied mathematics, applied statistics, engineering physics, physics, machine learning, or a related field, or have equivalent qualifications. Strong foundations in linear algebra, probability theory, calculus, programming, and machine learning are required. Additional qualifications in Bayesian statistics, mathematical modelling, probabilistic machine learning, deep learning, and large language models are meriting. Candidates should demonstrate creativity, thoroughness, structured problem-solving, and proficiency in English.

The position is fully funded as part of the DDLS Fellows program, with employment at 100% according to Uppsala University’s regulations. The starting date is 17 August 2026 or as agreed. The application deadline is 14 May 2026. Applicants should submit a cover letter, CV, degrees and transcripts, Master’s thesis or other scientific text, publications, references, and up to two reference letters. For further information, contact Assistant Professor Sara Hamis at [email protected].

Uppsala University is a broad research university with a strong international reputation, offering excellent working conditions and benefits. The position may be subject to security vetting. For more details about the DDLS program and working at Uppsala University, visit the provided links.