Developing secure applications for health providers and research-education (XtraTrack Season 2026)

27-05-2026 16:00

AI offers a significant potential in health care, not only in the analysis of multimodal health data and integration in interfaces and diagnostic assistants but also reducing the burden of documentation. The usage of existing AI solutions however can be difficult due to data protection guidelines and intransparent algorithms. “KISSKI” is one of four German BMFTR-funded AI service centers and focusses on critical infrastructures such as health care. Established as a partnership between leading universities, university medical centers, research institutions, and industry, KISSKI is uniquely positioned to address the critical challenges of deploying AI in sensitive and critical domains, with a primary focus on healthcare and energy. The core mission of KISSKI is to provide a highly available, secure, and scalable AI infrastructure that lowers the barrier to entry for both researchers and industry, particularly small and medium-sized enterprises (SMEs). We will demonstrate how KISSKI integrates cutting-edge research in scalable AI, heterogeneous hardware, secure data management, and automated model development into a comprehensive, user-friendly service portfolio. For researchers, our platforms provide access to powerful GPU and specialized hardware (FPGAs, neuromorphic chips) for training and inference, coupled with secure, compliant data environments. We support researchers through dedicated consulting, from initial concept evaluation to model deployment, ensuring their work meets the stringent requirements of critical infrastructure. For industry, KISSKI serves as a vital innovation partner. We offer a suite of services designed to accelerate the development and commercialization of AI solutions. This includes pilot project support, technical benchmarking, data management consulting, and specialized training in AI for healthcare and energy. Our goal is to help industry partners navigate the complexities of AI implementation, from data preparation to model validation and regulatory compliance (e.g., Medical Device Regulation), thereby reducing time-to-market and development costs. The lecture will highlight KISSKI's user-centered community approach that fosters collaboration between academia and industry. We will showcase real-world use cases from healthcare (e.g., AI for medical image analysis, predictive diagnostics) and administration (e.g., transferring voice notes into patient letters, optimizing workflows) illustrating how KISSKI's services are already driving innovation. Finally, we will discuss KISSKI's sustainable business model, ensuring long-term viability beyond initial funding. This includes a clear path to a self-sustaining, cost-recovery model for industrial projects, guaranteeing that the center remains a vital national resource for advancing AI in Germany's most critical sectors. KISSKI is not just a service center; it is a national ecosystem for responsible and impactful AI innovation.

## Prof. Dr. Julian Kunkel - Institut of Computer Science, Georg-August-University Göttingen; Gesellschaft für wissenschaftliche Datenverarbeitung, Göttingen

Prof. Dr. Kunkel is a Professor in High-Performance Computing at the University of Göttingen and the Deputy Head High-Performance Computing at GWDG (Gesellschaft für wissenschaftliche Datenverarbeitung Göttingen). Previously, he was a lecturer at the Computer Science Department at the University of Reading and a postdoc in the research department of the German Climate Computing Center (DKRZ). He manages several research projects revolving arounf AI, High-Performance Computing and particularly high-performance storage. Besides his main goal to provide efficient and performance-portable I/O, his HPC related interests are data reduction techniques, performance analysis of parallel applications and parallel I/O, management of cluster systems, cost-efficiency considerations, and the software engineering of scientific software. Since 2022, he is project coordinator of „KISSKI“ (AI Service center for critical infrastructures), one of four german AI service centers funded by the BMFTR. This service centers provide GPGPU ressources, LLMs and LLM-based services, training and consulting to all german institutions with minimal barriers of entry. As service center for critical infrastructures, KISSKI focusses on services with highest data protection, SecureGPC and Use Cases tailored to demands of clinics. Furthermore, KISSKI contributes to the establishment of an German AI network which unites providers of Open Source AI solutions. Prof. Dr. Kunkel is furthermore a member of many international program comittees, he is a founding member of the IO500 benchmarking effort, the Virtual Institute for I/O, and the HPC Certification Forum.

Organized by GMDS AG MoCoMed / Impressum / Privacy