Reliable Multi-Agent Systems for real-world use

Multi-agent systems offer a promising approach to automating complex tasks while providing flexibility across various applications. However, designing these systems to operate reliably in real-world conditions can be challenging. This workshop will explore the fundamental principles necessary for developing robust and scalable multi-agent solutions, while relying entirely on open source AI models and tooling. Conceptual Foundations Understanding the difference between free-form interaction (chat) and structured workflow automation. The importance of fault tolerance and monitoring in multi-agent systems. The role of open-source AI models in multi-agent systems and their advantages for transparency and customization. Hands-On with diskurs An introduction to the open-source framework diskurs, designed explicitly for building reliable multi-agent systems. A step-by-step implementation of a sample project, covering everything from setup and agent design to inter-agent communication. Testing and Validation An overview of a multi-agent testing environment based on open source frameworks. Strategies for automated testing, simulating real-world scenarios, and maintaining continuous quality control This workshop is suitable for anyone interested in AI-driven automation--whether you're a beginner with basic programming knowledge or an experienced developer looking to enhance your understanding of building and testing complex agent systems. By the end of the workshop, you will gain a clear conceptual overview and hands-on experience in creating robust multi-agent applications for your projects. Flurin Gishamer I am a Senior Data Scientist at Open Systems and the creator of Diskurs--an open-source framework for building multi-agent systems. My work spans natural language processing, large language models, and generative AI, and lately, I have been focusing mostly on multi-agent systems. I enjoy learning new things, exchanging ideas with others, and sharing my discoveries in ways that spark meaningful dialogue.