Research profile
Research & academic interests
Building a recommendation system for diagnosis support in emergency medicine — robust in complex, heterogeneous, and time-critical healthcare environments.
The main focus of my PhD at Christian-Albrechts-Universität zu Kiel is developing a recommendation system for diagnosis support in emergency medicine — a system that helps clinicians reach a diagnosis from incomplete, heterogeneous, and time-critical patient data. Around that goal I work on AI-supported integration of multimodal clinical data (text, image, tabular, and time-series), missing-value imputation, synthetic data generation, and the design and evaluation of the underlying deep-learning models. This work is grounded in the Intelligent Systems Group at Kiel University and in the KIT2 (TeleNotArzt) project — a completed real-world collaboration with RWTH Aachen University.
Current research focus
- A recommendation system for diagnosis support in emergency medicine (main PhD focus)
- Deep learning for heterogeneous and multimodal medical data
- Missing value imputation for medical time-series and tabular data
- Synthetic data generation for healthcare applications
- Disease classification and treatment recommendation in urgent care
Research strengths
- Design and evaluation of AI models for medical emergencies
- Preprocessing and imputation of incomplete clinical datasets
- Comparative analysis of deep learning and tabular generative models
- Multimodal learning across time-series, tabular, and text inputs
- Translating analytical results into decision-support tools
Research themes
Research approach
I approach research through rigorous modeling, comparative evaluation, and practical relevance to real-world healthcare problems. The goal is a diagnosis recommendation system that is reliable under the conditions of emergency care — missing values, mixed data types, and decisions made under time pressure — by combining advanced machine-learning methods with a domain-specific understanding of clinical data quality and uncertainty.