The AI topic came up in both MicroProfile and Jakarta EE meetings. The MicroProfile AI group was set up in 1Q 2024 and the group has made significant progress towards writing MicroProfile and Jakarta EE AI-infused applications connecting with a LLM via configuration. The weekly discussions and recordings can be accessed
here.
The MicroProfile AI group evaluated some Java AI libraries and also attempted to create a new set of APIs. After some investigation and prototype, the MicroProfile AI group decided to adopt the defacto standard Langchain4J and concentrated on the integration of Langchain4J with Jakarta CDI, MicroProfile Fault Tolerance, MicroProfile Telemetry, etc. In this way, the Java communities can work together and put wisdom in one place. You can see the prototype
here. As part of the prototype, a sample application was developed to work with some popular runtimes including Liberty, Quarkus, Helidon and Glassfish. You can try out the samples from
here.
The prototype is nearly wrapped up. Now it is time to discuss a potential home for the integration.
Option 1: contribute to the codebase of
langchain4j and place it under a repo like langchain4j-microprofile-jakarta, leveling up with langchain4j-spring. Both Jakarta and MicroProfile communities can freely contribute and promote this work.
Option 2: place this integration under the MicroProfile github repo as a library.
It seems
Option 1 has a greater advantage. We will discuss this further at the next week's MP AI call on Monday. The call details are in the
MP Calendar. Please join the discussion if you would like to share your point of view. Please comment if you have any thoughts on this.
Thanks
Emily
-- 