The Department of Diagnostics and Intervention invites applications for a two‑year postdoctoral fellowship within the project AI‑Driven Medical Imaging. The position is expected to start on 1 September 2026 or as agreed. The fellowship is funded by The Kempe Foundations.
Project description
Machine learning and artificial intelligence have had a major impact on medical image analysis in recent years. While CT and MRI provide highly standardised and robust diagnostic imaging at high cost, ultrasound enables portable and cost‑effective applications. A key challenge, however, is the strong dependence on operator expertise, both during data acquisition and clinical interpretation. The aim of this postdoctoral project is to develop and evaluate AI‑based methods that reduce operator dependence in ultrasound imaging, with a focus on clinically relevant applications. The research is conducted within a well‑established research group led by Christer Grönlund, specialising in methodological development in medical ultrasound imaging.
The work encompasses two main application areas:
1) Atherosclerosis diagnostics Atherosclerosis is a common cardiovascular disease that may lead to plaque rupture and severe conditions such as stroke. Current diagnostic methods have limited ability to identify patients at intermediate risk, despite this group accounting for a substantial proportion of disease burden. The project focuses on early identification of pathophysiological phenotypes, decentralised examinations, and reduced operator dependence. 2) Diagnostics of skeletal muscle diseases In a parallel research track, non‑invasive ultrasound‑based methods are developed as alternatives to current needle‑based diagnostic techniques. The work includes advanced analysis of motor units in muscle tissue, with the goal of translating methods from specialised research systems to clinically available ultrasound scanners.
Within the postdoctoral project, the fellow will have the opportunity to further develop and deepen expertise in methods and concepts such as:
Anomaly detection
Domain adaptation
Few‑shot learning
Graph‑based models
Vision transformers and/or diffusion models
2D+time (video) segmentation
Qualifications To be eligible for a postdoctoral fellowship, the applicant must have been awarded a doctoral degree, or a foreign qualification deemed equivalent to a PhD, in biomedical engineering, computer science, electrical engineering, or a related field. The degree requirement must be fulfilled no later than the time of the fellowship decision.
Priority will be given to applicants who obtained their doctoral degree no more than three years prior to the application deadline. Exceptions may be made in cases of documented special circumstances, such as sick leave, parental leave, clinical duties, union assignments, or similar circumstances.
Additional requirements
Documented education and practical experience in machine learning and AI methods, demonstrated through coursework and/or scientific publications.
Excellent written and spoken English skills.
Strong programming skills in Matlab and/or Python.
Meritorious qualifications
Experience in medical image analysis. Knowledge of or practical experience with generative models, anomaly detection, and image and signal analysis.
Personal qualities We are looking for a motivated and curious candidate with strong collaboration skills, who is also able to work independently and take responsibility for their research.
Application The application must be written in English or Swedish, and all documents should be submitted in Word or PDF format. Applications must be registered via Umeå University’s e‑recruitment system Varbi and received no later than 30 April 2026.
