MLOps & LLMOps in Banking IT: Automation, Infrastructure & Real-World Challenges

To bring the power of AI into everyday banking operations, robust infrastructure for MLOps and LLMOps is essential—especially in highly regulated environments. At Finanz Informatik, the digitalization partner of the Sparkassen-Finanzgruppe, teams have a long-standing track record of building secure, scalable, and automated platforms for mission-critical banking solutions. This foundation is now being extended into the AI era. In this session, Sebastian, Daniel, Lars and Matthias share their best practices and lessons learned from operating AI solutions inside air-gapped banking infrastructures. They walk you through key components like model registries, experiment tracking, model monitoring, and the automated provisioning of LLMs—including real-world challenges from inference server operations. The team will also present the S-AI-Pilot, a generative AI assistant already in use at Sparkassen workplaces. What does it take to build enterprise-grade MLOps and LLMOps under the toughest constraints? Join us for technical insights and practical guidance from the core team behind one of the largest AI initiatives in the German banking sector.