and
Medical University of Graz
Agentic AI for Multimodal Medical Data Analysis
Objective:
Recent advances in multimodal foundation models enable the joint analysis of medical images, clinical text, laboratory values, signals, and other patient information. However, directly processing all available data with a single model can limit reliability and the ability to perform specialized quantitative analyses.
The objective of this thesis is to investigate an agentic AI framework for multimodal medical data analysis, in which a reasoning agent dynamically plans an analysis workflow and coordinates specialized models and computational tools. The work will explore existing (medical) agent architectures, implement a prototype system, and evaluate different strategies for planning, tool use, multimodal information integration, and the generation of clinically meaningful outputs.
Further Information:
- Z. Wang et al., “MedAgent-Pro: Towards Evidence-Based Multi-Modal Medical Diagnosis via Reasoning Agentic Workflow,” ICLR, 2026.
- Y. Du et al., “CARE: Towards Clinical Accountability in Multi-Modal Medical Reasoning with an Evidence-Grounded Agentic Framework,” ICLR, 2026.
Qualifications:
- Good Python programming skills and experience with deep learning frameworks such as PyTorch
- Interest in multimodal AI, AI agents, and medical applications
- Experience with LLMs/VLMs, tool-using agents, or medical image analysis is advantageous
Contact:
Darko Stern
darko.stern@medunigraz.at
Martin Urschler
martin.urschler@tugraz.at