Real-World AI Patterns with Spring AI and Vaadin by Marcus Hellberg / Thomas Vitale @ Spring I/O 25

Spring I/O 2025 - 21-23 May, Barcelona Slides: https://speakerdeck.com/thomasvitale/... For the past few years, we have been experimenting with large language models (LLMs) and have discovered effective ways to integrate AI capabilities into Java applications. But how do we transition from experimentation to actual production adoption? This talk will analyze the most common AI patterns used in real-world applications, complemented by live demos showcasing their implementation with Spring AI and Vaadin. You will learn how to: Provide short-term and long-term memory to a large language model (LLM). Integrate LLMs with your APIs through tool calling and build advanced agents. Protect your AI workflows from prompt injection attacks and sensitive data leaks using guardrails. Interact with models via text, image, audio, and video through multimodality. Enhance the context provided to LLMs from your data using advanced retrieval techniques. You will gain practical insights into how these patterns can be effectively integrated into your projects, with examples you can implement in your existing applications immediately, whether you are running your workloads on-premises or in the cloud.