Postdoc Researcher AI/Machine Learn for Multimodal Cardiovasc Data (VASCUL-AID)

VASCUL-AID is a €6.3 million Horizon Europe project (2023–2029), coordinated by Amsterdam UMC, that brings together 16 partner organizations across Europe. Its goal is a trustworthy AI platform that predicts cardiovascular risk and disease progression in patients with abdominal aortic aneurysm (AAA) and peripheral artery disease (PAD). Both conditions are common, often silent until late, and carry a high risk of major cardiovascular events, amputation and death.

A key challenge in treating these patients is that it is still unclear which risk factors drive disease progression and the development of other cardiovascular conditions. The consortium therefore collects and integrates rich longitudinal patient data from major European clinical centres: electronic health records, proteomics, genomics, imaging (CT/MRI), wearable sensor data and clinical follow-up. These data form the basis for models that support personalized risk assessment and treatment decisions.

More information about the project: www.vascul-aid.eu. 

As a postdoctoral researcher, you will lead the AI/machine-learning and multimodal data-integration work of VASCUL-AID at Amsterdam UMC. You will design, build and validate models that predict disease progression and clinical outcomes in AAA and PAD patients, using the full breadth of the consortium's data.

Your work spans the whole modelling pipeline: from harmonizing heterogeneous clinical and molecular data, to developing graph-based and foundation-model approaches, to delivering robust, uncertainty-aware models that clinicians can trust and use. You will also drive data-driven biomarker discovery together with biomedical and clinical colleagues. You will be responsible for governance of the data data, handling data (harmonization, missing data, synthetic data), coordinating data collection together with the team. You will be helping in supervising the consortium tasks and PhD-students working on prediction models.

You will be embedded in the Department of Vascular Surgery and work closely with vascular surgeons, clinical researchers, data scientists and partner institutions across Europe. This position offers a unique opportunity to translate advanced machine learning and multi-modal prediction models into real clinical impact, in a highly visible international consortium.

• Lead the development of machine-learning models/multi-model prediction models that predict cardiovascular disease progression and outcomes in AAA and PAD patients.
• Integrate multimodal patient data: EHR, proteomics, genomics, imaging, wearables and clinical data into coherent analysis pipelines and representations.
• Apply and advance state-of-the-art methods, including graph representation learning (e.g. graph neural networks), foundation-model embeddings, and multi-omics integration.
• Develop uncertainty-aware models and rigorous validation strategies (internal and external validation, calibration, robustness across cohorts).
• Contribute to biomarker discovery by identifying predictive molecular and clinical signatures, in collaboration with biomedical and clinical partners.
• Prepare models for deployment: reproducible code, documentation, model packaging, and integration with consortium tools and clinical workflows.
• Coordinate and governance of the data collection, handling of data (work with missing data, data harmonization process and creating synthetic data)
• Coordinate the AI/ML work package activities at Amsterdam UMC and act as the technical point of contact for consortium partners.
• Supervise and support PhD candidates, master students, and junior data scientists.
• Publish in high-impact journals and conferences, present at international meetings and contribute to consortium reports and deliverables.
• Contribute to follow-up grant proposals and new research lines in AI for vascular medicine.

Qualifications

We are looking for an enthusiastic, independent researcher who wants to make a difference for patients with vascular disease. You have:

• A completed PhD in machine learning, computer science, computational science, data science, biomedical engineering, bioinformatics or a closely related field.
• Strong expertise in machine learning and deep learning, preferably including graph neural networks, foundation models, and representation learning.
• Demonstrable experience with biological, clinical and/or omics data (e.g. EHR, proteomics, genomics, imaging or wearable data).
• Excellent programming skills in Python and hands-on experience with PyTorch (or a comparable framework).
• Experience working with high-performance computing (HPC) environments and GPU-based training.
• A track record of peer-reviewed publications in AI/ML or medical informatics venues (e.g. NeurIPS, ICLR, MICCAI) and/or translational journals (e.g. npj Digital Medicine).
• Good communication skills and the ability to explain complex methods to clinicians and non-technical audiences.
• Fluency in English, spoken and written. Knowledge of Dutch is a plus.
• The ability to work in a cross-functional, international consortium, with a proactive and collaborative attitude.
• Experience with uncertainty quantification, survival or time-to-event modelling, MLOps / model deployment, or cardiovascular research is an advantage.