Large language models enable the automation of many complex tasks and open up new possibilities for knowledge-based applications and user interface design. However, the high specificity and speed of development often lead to significant knowledge gaps and a partially unrealistic understanding among many decision-makers. This can easily result in poor decisions and inappropriate applications. Architects must act in a timely manner and hence need a solid understanding of the technology, as well as its possibilities and limitations.

This advanced course introduces the basic concepts of generative AI. The focus is on Large Language Models (LLM) such as GPT, DeepSeek, or LLAMA, and the applications and architectures based on them (especially Retrieval Augmented Generation, RAG). Subsequently, participants will have the opportunity to discuss AI projects planned within their organization from an architectural perspective (appropriateness of the technology) and value creation perspective and to identify pitfalls. The course covers the following topics in detail:

• Fundamentals of LLMs (tokens, embeddings, transformers, etc.)
• Use of commercial and open-source LLMs (APIs, libraries, prompt engineering)
• Application areas for AI/LLM-based applications (classification, topic modeling, semantic search, multimodal applications, etc.)
• Architectures for AI/LLM-based applications (especially RAG)
• Security aspects (including possible solutions)
• Approaches to evaluation and management from an EA perspective

1 day