Developing AI applications in Java with Quarkus and Langchain4j

Speaker: Jan Martiška (Red Hat, https://x.com/janmartiska, https://www.linkedin.com/in/janmartiska/)

Abstract:

In this practical session, I will demonstrate how Quarkus, the cloud-native Java development framework, and its Langchain4j extension can be used for working with AI models like GPT and others, including image-generating models, opening the door to endless possibilities.
I will show how to use Quarkus to:

• Build a chatbot application that allows you to supply data from your own custom data store and feed it to the AI model, enabling it to answer questions using this data (this technique is called Retrieval Augmented Generation).
• Build a highly autonomous agent by supplying the model with various tools, that means locally implemented functions that the model can decide to execute when necessary (for example, send something via email, or write to a database), and then letting the agent determine the appropriate sequence of steps to accomplish the given high-level goal using these tools.
• Demonstrate the Quarkus goodies that make the development process easier, like the interface for manually chatting with the model directly through the application and for exploring image-generating capabilities of image models.

The development joy of Quarkus with all its goodies and incredible productivity now also extends to creating AI applications. Join me and see for yourself!

--
What to say? This talk probably meets all current criteria. Java? Check. AI? Check. One of the best current frameworks? Check. The practical applicability of AI in your workload is surely an interesting topic that we need to learn about!

We will also stream this time online at https://www.youtube.com/watch?v=gQMue4bMp1g! If you plan to attend online, you don't need to register here since we're using this information to order food. This is also a small push for you to come in person!

After the talk, we will have the after-party as usual.