GenAI & BeyondHands-on Lab180min
From Scratch to Scalable: Building Smarter AI Agents with Frameworks
This workshop guides Java developers from building a basic AI agent with manual LLM integration and memory management to transforming it into a scalable, production-ready solution using agent frameworks, offering hands-on insight into both core agent components and advanced framework capabilities.
Jettro CoenradieLuminis
Daniël SpeeLuminis
talkDetail.whenAndWhere
Tuesday, October 7, 09:30-12:30
BOF 1
Detailed Description
It doesn’t take much to build a simple AI agent — with just basic programming skills, you can create one from scratch without using any frameworks. But moving from a working prototype to a production-ready agent introduces a whole new level of complexity.
In this hands-on workshop, you’ll start by building your own AI agent in Java. You’ll implement LLM integration, tool calling, and manage conversational history — all from the ground up. This gives you a deep understanding of how agents operate under the hood.
With that foundation, you’ll then transition to using an agent framework. You’ll learn how frameworks simplify the development process by handling complexity like multi-LLM support, observability, memory, and deployment. By the end, you’ll have transformed your home-grown agent into one that’s framework-ready and production-capable.
Key Takeaways:
Target Audience:
Java developers who are curious about AI agents and want hands-on experience creating, understanding, and scaling them for real-world applications.
Joining the workshop
Follow the link to the repository to prepare for the workshop.
It doesn’t take much to build a simple AI agent — with just basic programming skills, you can create one from scratch without using any frameworks. But moving from a working prototype to a production-ready agent introduces a whole new level of complexity.
In this hands-on workshop, you’ll start by building your own AI agent in Java. You’ll implement LLM integration, tool calling, and manage conversational history — all from the ground up. This gives you a deep understanding of how agents operate under the hood.
With that foundation, you’ll then transition to using an agent framework. You’ll learn how frameworks simplify the development process by handling complexity like multi-LLM support, observability, memory, and deployment. By the end, you’ll have transformed your home-grown agent into one that’s framework-ready and production-capable.
Key Takeaways:
- Understand the core components of an AI agent by building one manually.
- Learn how to integrate LLMs, tools, and memory in Java.
- Gain experience with agent frameworks to support scalability and production use.
Target Audience:
Java developers who are curious about AI agents and want hands-on experience creating, understanding, and scaling them for real-world applications.
Joining the workshop
Follow the link to the repository to prepare for the workshop.
Jettro Coenradie
Jettro is a software architect, search relevance geek, and Gen AI enthusiast who loves to talk about his job, hobbies, and other things that inspire people. Jettro truly believes in the Luminis mantra that the only thing that grows by sharing is knowledge. After more than ten years of creating the best search engines for multiple customers, Jettro is drawn into AI, LLMs and Agents. Learning and talking about these technologies is what drives him to keep improving the user experience of information-driven solutions.
Daniël Spee
Daniël is a software engineer at Luminis with a passion for search, AI, and data-centric systems. He believes that great search is more than just technology: it’s about understanding user intent, the structure and meaning of data, and the business processes behind it. Daniël combines AI with classical search techniques to build smarter, context-aware systems. His current focus is on how AI agents and retrieval pipelines can automate and improve real-world workflows.
comments.speakerNotEnabledComments