3F Ebisu Neonato, 4-1-18 Ebisu, Shibuya-ku, Tokyo
The first study session of the Japan Jakarta EE & MicroProfile User Group will cover CDI and AI utilization.
CDI is the dependency injection mechanism standardized in Jakarta EE servers. Its full name is Jakarta Contexts and Dependency Injection, and its defining feature is that it incorporates Contexts, not just Dependency Injection. This makes it slightly more challenging to use compared to Spring's DI. For this session, Mr. Yoshizaki from Accenture will give a lecture on practical CDI usage. Using CDI merely as a DI technology would be a waste, so please join us.
The latter half will cover AI integration, which is expected to be standardized in Jakarta EE/MicroProfile in the future. Following the shift from EJB-oriented to CDI-oriented approaches in Jakarta, AI integration is also being explored based on CDI. While Python, geared towards data scientists, has been the mainstream programming language for AI use until now, Jakarta EE application developers can finally integrate AI into their own applications using the CDI-oriented Jakarta style.If you're interested in using AI with Jakarta EE servers, please look forward to this session.
We also plan to hold a networking session. Light refreshments and drinks will be provided, so please join us if you have time.
※ Light refreshments will be served during the networking session. Participation is free of charge.
| Time | Content |
|---|---|
| 19:00–19:05 | Opening |
| 19:05–19:55 | Common Misunderstandings of CDI Specifications and Tips for Handling Them |
| 19:55–20:05 | Break |
| 20:05–20:55 | LangChain4j-CDI: A Different Flavor from LangChain4j |
| 8:55 PM - 9:00 PM | Closing |
| 9:00 PM - 10:00 PM | Networking Reception |
In Jakarta EE development, the DI specification Jakarta Contexts and Dependency Injection (CDI) serves as the ecosystem framework connecting components.When assigned to a CDI project, it's easy to fall back on experience with other DI frameworks to understand CDI. However, CDI has subtle differences. I'll share what these differences are, based on actual problems I encountered and their solutions in CDI projects. I'll also run a simple demo (source code to be released later) on JBoss EAP, demonstrating these issues while explaining them.I hope this serves as useful background knowledge for those about to start using CDI.
Futoshi Yoshizaki (LinkedIn)
Accenture Japan Ltd. Accenture Technology IT Solution
Slides:
https://speakerdeck.com/futokiyo/cdinowu-jie-sigatinashi-yang-tosonodui-chu-tips
https://github.com/jjugnsjeecditips202602
While several libraries exist for handling LLMs from Java, LangChain4j is currently one of the most widely used. This session introduces not just LangChain4j itself, but LangChain4j-CDI—LangChain4j integrated with Jakarta CDI (Context and Dependency Injection).
The origin of LangChain4j-CDI emerged during the process of attempting to create "MicroProfile AI (tentative name)". Following discussions within the MicroProfile community, it was ultimately decided not to create something new within MicroProfile, but rather to contribute to LangChain4j.
Although named "CDI," LangChain4j-CDI integrates with various Jakarta EE and MicroProfile technologies. It provides a familiar interface for Jakarta EE (Java EE) / MicroProfile users, enabling them to create Jakarta EE applications that interact with AI systems.
Additionally, this session will briefly cover AI trends related to Jakarta, not limited to LangChain4j-CDI.
Kenji Kazumura (@kkzr)
Fujitsu Limited / Jakarta EE Specification Committee Member
Slides:
https://speakerdeck.com/kazumura/langchain4jtoha-wei-wei-ulangchain4j-cdi
https://github.com/fujitsu/app_blog/tree/master/lc4j-cdi-samples
This is a user group for Jakarta EE and MicroProfile in Japan. Enterprise Java technology, which began with Java EE (J2EE), has now been renamed Jakarta EE and MicroProfile. It continues to evolve...
Join community
