Postdoctoral Research Assistant in Machine Learning and Multimodal Clinical AI

We are seeking two ambitious postdoctoral machine-learning researchers to join the Spatial and Translational Pathology Group at the University of Oxford. This 24-month post will develop advanced AI models using a unique, large-scale collection of multimodal NHS clinical data, including electronic health records, laboratory and molecular data, medical imaging and digital pathology.

The posts form part of a multidisciplinary programme in computational pathology and clinical AI supported by the MPN Research Foundation, Blood Cancer UK, Cancer Research UK and Wellcome. The successful candidates will develop and validate advanced machine-learning approaches using large-scale multimodal clinical, molecular and imaging data, with applications including personalised prediction of disease progression, treatment response and patient outcomes.

One of the posts will have a particular focus on a major MPN Research Foundation programme establishing a multimodal NHS data resource for approximately 5,000 patients with, or investigated for, myeloproliferative neoplasms (MPNs).

We are particularly interested in candidates with a strong machine-learning, computer-science or engineering background. Prior experience in haematology or pathology is not essential: the successful candidate will work closely with clinician scientists, pathologists, biomedical engineers and computational researchers and will be supported to develop the necessary clinical and biological understanding.

• Lead the design, development and validation of machine-learning models integrating multimodal clinical, laboratory, molecular/genomic, imaging and pathology-derived data.
• Develop robust and reproducible pipelines for large-scale, heterogeneous and longitudinal NHS data, including data representation, harmonisation, quality control and linkage within approved secure environments.
• Work closely with the Thames Valley and Surrey Secure Data Environment (TVS SDE) team, with shared learning, knowledge exchange and training to develop secure, reproducible approaches for machine-learning research using large-scale NHS clinical data.
• Develop and evaluate appropriate multimodal learning approaches, including methods for incomplete and temporally structured data, with rigorous comparison against relevant clinical and statistical baselines.
• Work with multidisciplinary collaborators to translate clinical questions into tractable computational problems and interpret model outputs in their biological and clinical context.
• Contribute machine-learning expertise to other closely aligned projects within the research group, including reusable code and pipelines, technical advice, and support for research students and colleagues using AI/ML methods.
• Contribute to high-quality publications and present findings at national and international scientific meetings.

You will hold, or be close to completion of, a PhD/DPhil in machine learning, computer science, engineering or a related quantitative discipline. You should have:

• Strong programming skills in Python.
• Experience developing and evaluating machine-learning or deep-learning models.
• Experience working with large, complex and/or multimodal datasets.
• Ability to work effectively and communicate within a multidisciplinary research team.
• Experience with clinical or healthcare data, multimodal machine learning, biomedical AI, digital pathology or secure data environments would be advantageous but is not essential.

The postholder will join the Spatial and Translational Pathology Group within the Nuffield Division of Clinical Laboratory Sciences, Radcliffe Department of Medicine. The programme brings together expertise in clinical haematology, pathology, biomedical engineering, machine learning and data science, with access to large-scale NHS data and University high-performance CPU/GPU computing infrastructure. The postholder will work in close partnership with the Thames Valley and Surrey SDE team and will benefit from shared learning and training in secure, large-scale NHS data research.