Data & AIConference45min
Building full-stack AI agents: From project generation to code execution
This session explores advanced architectural patterns for creating production-ready AI agents, emphasizing multi-phase reasoning, structured prompt management, and incremental development. Through real-world examples, it demonstrates building maintainable, reliable agents, featuring a live demo where AI agents generate and deploy a web application using prompts. Ideal for developers and architects.
Stephan JanssenDevoxx
talkDetail.whenAndWhere
Thursday, April 17, 11:35-12:20
Neuilly 251
We'll delve into proven architectural patterns for building production-ready AI agents, moving beyond basic prompt engineering to focus on the essential components of robust systems. These include multi-phase reasoning with distinct planning and execution stages, structured prompt management with version control, and the adoption of incremental development practices. Through real-world code examples, we'll demonstrate how these patterns enable the creation of maintainable and reliable agents. This talk is perfect for developers and architects aiming to implement AI systems capable of consistently handling complex workflows. As part of the session, I'll showcase a live demo where AI agents generate and deploy a fully functional web application using prompts alone.
Stephan Janssen
As an organizer of Devoxx Belgium, I am passionate about conducting research and development (R&D) that leads to new features in Devoxx-related applications like the CFP web app. In recent years, I have been exploring and experimenting with cutting-edge technologies such as large language models (LLMs), convolutional neural networks (CNNs), and other artificial intelligence (AI) tools. Recently I've been working on the DevoxxGenie IntelliJ plugin which allows you to talk to LLM's both locally and remotely.
comments.speakerNotEnabledComments