Data Scientist/Computational Scientist – Multimodal and Multi-Agent AI for Precision Oncology (m/f/d)

Universitätsklinikum Heidelberg; Heidelberg; Vor Ort; EUR 70.000 - 100.000; Vollzeit

Zusammenfassung

Universitätsklinikum Heidelberg is seeking a researcher to lead AI development for a multimodal, explainable decision-support system in prostate cancer care. You will integrate longitudinal clinical data, imaging, pathology and outcomes to model patient trajectories and optimize treatment selection.

The role requires strong Python/R coding, hands-on ML expertise in deep learning and survival analysis, with experience in multimodal datasets and collaboration across international partners.

Qualifikationen

• Master's degree or PhD in Data Science, Bioinformatics, CS, Computational Biology or related discipline.
• Strong programming skills in Python and/or R.
• Expertise in machine learning, deep learning and predictive modelling.
• Experience with multimodal, longitudinal or high-dimensional datasets.
• Interest in multi-agent systems, agent orchestration, transformers or clinical decision support.

Aufgaben

• Integrate and preprocess multimodal longitudinal clinical data, imaging, pathology and outcomes.
• Develop ML components for patient-trajectory modelling, recurrence and survival prediction.
• Contribute to architecture for explainability, uncertainty estimation and human-in-the-loop mechanisms.
• Perform internal and external validation, benchmarking against risk models across European cohorts.
• Collaborate with clinicians, imaging experts and AI researchers for publications and knowledge transfer.

Kenntnisse

Python

ML

Deep learning

Multimodal data

Explainable AI

Survival analysis

Reproducible research

Ausbildung

Master's or PhD in Data Science / related field

Tools

Python

R

Jobbeschreibung

Join the PROMETHEUS consortium, a four-year EIC Pathfinder project developing a next-generation multimodal, explainable and multi-agent AI system for prostate cancer diagnosis and treatment selection (https://cordis.europa.eu/project/id/101306686). The consortium brings together leading European partners with multicentre patient cohorts, including Heidelberg University Hospital, DKFZ, Erasmus MC Rotterdam, Radboudumc Nijmegen, Karolinska Institutet and Umeå University.Within the group of PD Dr. Magdalena Görtz (https://www.dkfz.de/en/multiparametric-methods-for-early-detection-of-prostate-cancer), you will lead the development of an AI agent for treatment selection that combines longitudinal clinical data, imaging, outcome prediction and knowledge-based reasoning. The goal is to move beyond isolated prediction models towards an integrated agentic system that can model patient trajectories, compare treatment strategies and generate transparent, clinically meaningful decision support.

Your Responsibilities

• Multimodal data integration and preprocessing: Harmonize and integrate longitudinal clinical data, imaging, pathology and outcome data
• Multi-agent and predictive AI development: Develop machine-learning components for patient-trajectory modelling, recurrence and survival prediction, and integrate them into the PROMETHEUS multi-agent
• Agent interaction, explainability and uncertainty: Contribute to architectures that combine specialized AI components, knowledge-based reasoning and clinically interpretable outputs; implement uncertainty estimation and human-in-the-loop mechanisms
• Validation and clinical translation: Perform internal and external validation, benchmark against established risk models and assess clinical utility across European cohorts
• Scientific communication and collaboration: Contribute to high-impact publications, presentations and close collaboration with clinicians, imaging experts and AI researchers in Heidelberg and across the European consortium

Your Profile

• Master's degree or PhD in Data Science, Bioinformatics, Computer Science, Computational Biology, Biostatistics, Medical Informatics, Mathematics, Physics, or a related discipline
• Excellent programming skills in Python and/or R
• Strong expertise in machine learning, deep learning and predictive modelling
• Experience working with multimodal, longitudinal or high-dimensional datasets
• Interest in or experience with multi-agent systems, agent orchestration, transformers, multimodal fusion, knowledge-grounded AI or clinical decision support systems
• Sound understanding of model validation, calibration and reproducible research practices
• Experience in explainable AI, survival analysis, medical imaging or healthcare AI is considered an advantage
• A proactive, independent and scientifically motivated mindset combined with excellent communication and collaboration skills in an interdisciplinary international environment

About us

For further information, please contact PD Dr. Magdalena Görtz, Senior Physician and Research Group Leader.

Advance the Future of Precision Oncology