Who is this training for?
Interested parties who wish to acquire further competencies in this topic. Basic prior knowledge in human-centered methods and advanced knowledge in Prompt Engineering, Context Engineering, RAG, Knowledge Graphs, AI automation, and low-code are required.
Why this training
More and more logic sequences and process automations are being taken over by AI agents. These processes often happen silently, and risks and errors remain unnoticed. It is therefore necessary to understand how to stabilize agentic workflows and make them reproducible – but most importantly how to bring humans back into control while minimizing token costs.
Contents
- Introduction to Large and Small Language Models
- Presentation of a simple agentic architecture
- Deep dive into Prompt Engineering requirements, e.g., regarding agent "handoffs"
- Deep dive into the Model Context Protocol (MCP)
- Use of state machines for multi-agent systems
- Presentation of the BMAD stack
- Practical examples of multi-AI agents (e.g., for generating entire software or similar)
Methodology
- Alternation between theory, demonstration, and practical sections
- Practical sections alternate between individual and group tasks
- Online workshop in FigJam, n8n, VS Code, Git, Obsidian, and other tools