About this role
Do you want to develop deep learning methods to create digital twins of coronary arteries? And help interventional cardiologists make smarter, data-driven treatment decisions in real time? You will contribute to a translational research project aimed at improving the diagnosis and treatment of coronary artery disease and supporting clinical decision-making during catheterization procedures. The project brings together cardiovascular imaging, generative AI, and computational modeling to develop probabilistic digital twins of coronary arteries. The goal is to create patient-specific models that combine anatomical and hemodynamic information derived from routine clinical imaging. By providing quantitative insights into coronary anatomy, blood flow, pressure, and wall shear stress—together with information about model uncertainty—the project aims to support more precise, physiology-guided treatment decisions without the need for additional invasive measurements. The research is embedded within the Quantitative Healthcare Analysis (qurAI) group and conducted in close collaboration with the CARA Lab and clinical partners in the Netherlands and abroad. You will work with large, multi-center clinical datasets and contribute to translating advanced computational methods into clinically relevant applications. This PhD position offers a unique opportunity to work at the intersection of artificial intelligence, cardiovascular engineering, and clinical medicine, with the potential to contribute directly to the future of personalized cardiovascular care. As a PhD candidate, you will work on improving the diagnosis and treatment of coronary artery disease, with a clear focus on clinical application during catheterization procedures. Your research sits at the intersection of medical image analysis, generative AI, and blood flow modeling. You will: develop deep generative models (such as latent diffusion models, implicit neural representations, and flow matching) for uncertainty-aware 3D reconstruction of coronary anatomy from 2D X-ray angiography; develop physics-informed neural networks and graph-based neural operators for fast estimation of 3D coronary hemodynamics (velocity, pressure, and wall shear stress fields); design computational pipelines that integrate image-based anatomy, blood flow physics, and uncertainty quantification; validate the developed methods on retrospective multi-center clinical datasets, in collaboration with clinical partners in the Netherlands and abroad; critically assess model performance and interpret results in the context of clinical and hemodynamic relevance; write scientific articles for high-impact journals and present your findings at national and international conferences and workshops. You are a motivated researcher with a strong technical background and a genuine interest in applying artificial intelligence to real clinical problems. You are curious, creative, and rigorous: you enjoy developing new methods and you care deeply about thorough validation and their impact on patient care. You have: an MSc degree in biomedical engineering, applied mathematics, computer science, physics or a related discipline; demonstrable experience with scientific programming (preferably in Python, including deep learning frameworks such as PyTorch); affinity with medical image analysis and computational modeling; experience with neural networks for image analysis, generative modeling, geometric deep learning or physics-informed machine learning, or you are willing to learn these quickly; strong collaboration skills: you enjoy working in a multidisciplinary team and feel comfortable interacting with clinicians; excellent English communication skills (speaking and writing); only applications in English will be considered. Experience with cardiovascular imaging or computational fluid dynamics is an advantage. A flying start to your career in scientific research, with the opportunity to obtain a PhD degree.