This research presents a reference architecture for organizations seeking to implement LLMs while maintaining data protection compliance and security.

• Addresses the critical challenge of data confidentiality risks in cloud-based LLM solutions
• Proposes a closed LLM system that meets regulatory requirements without sacrificing performance
• Offers an open-source framework that organizations can adapt to their security needs
• Balances technical capabilities with legal compliance demands

This architecture matters for Security professionals by providing a structured approach to implement LLMs that protect sensitive data while enabling AI capabilities within regulated environments.

Datenschutzkonformer LLM-Einsatz: Eine Open-Source-Referenzarchitektur