PhD Candidate in Explainable AI and Foundation Models for CT Imaging

Descrizione

Testo originale dell'annuncio (in inglese)

Welcome to Maastricht University! Do you want to contribute to advancing AI in medical imaging? In this PhD position, you will conduct research on explainable artificial intelligence and foundation models for CT imaging. You will develop and evaluate novel methods, with a strong focus on methodological innovation, rigorous validation and clinical relevance. PhD Candidate in Explainable AI and Foundation Models for CT Imaging Our goal: To develop novel AI methods for explainability in medical imaging, including diffusion-model-based approaches, and to advance foundation models for CT through model development, training, evaluation, and external validation. Your colleagues: You will join the Department of Precision Medicine at Maastricht University, embedded within GROW and the Faculty of Health, Medicine and Life Sciences. You will work in an interdisciplinary environment connecting artificial intelligence, medical imaging, and clinical translation. What you do As a PhD candidate, you will undertake a four-year doctoral research project leading to a PhD thesis. You will develop and evaluate new methods for explainable AI in medical imaging, with particular attention to the use of diffusion models for explanation and interpretation. You will also contribute to the design, training, adaptation, and external validation of foundation models for CT images. Are you ready to set the course for the years ahead? Then we’d love to meet you. What you bring We’re not looking for checkboxes; we’re interested in who you are and what you bring. Do you recognize yourself in this? You are an analytical and curious researcher with an interest in technically innovative research at the intersection of artificial intelligence, medical imaging and clinical translation. You enjoy tackling complex problems and working in an interdisciplinary and international environment, while taking ownership of your work and developing as an independent researcher. You approach research systematically and have experience developing well-structured, well-documented and reproducible research code and organising experiments in a way that enables results to be reproduced and your work to be understood and further developed by others. You are motivated to further develop as an independent researcher and successfully complete your PhD within the appointment period. Furthermore, you bring: You hold, or will shortly obtain, a Master’s degree in Artificial Intelligence, Computer Science, Biomedical Engineering, Medical Image Analysis, Applied Mathematics, Data Science, or a closely related field. You have a solid theoretical and practical background in machine learning and deep learning, including experience developing, training and evaluating models, preferably for image analysis tasks. You have strong programming skills in Python, including the ability to develop and adapt code for machine-learning experiments, train and evaluate deep-learning models, and process and analyse experimental results. You have hands-on experience using a deep-learning framework, preferably PyTorch to develop and adapt code, train deep-learning models, and evaluate their performance. You have knowledge of, or a strong interest in, the principles of generative modelling and an interest in applying and further developing generative approaches, including diffusion models, for medical imaging. You have a C1 level of proficiency in written and spoken English, according to the Common European Framework of Reference for Languages (CEFR). The following qualifications are considered advantageous: Experience with explainable AI, uncertainty estimation, trustworthy AI, or model interpretability. Experience with generative models, particularly diffusion models. Experience with foundation models, self-supervised learning, representation learning, or large-scale pretraining. Experience with medical imaging, particularly CT, and associated image formats or processing workflows. Familiarity with DICOM, NIfTI,