I Member of Jakarta EE Specification Committee Member of JCP Executive Committee Board of Director of Eclipse Foundation Member of Adoptium Steering Committee 3
AI Predictive Generative Artificial Intelligence Machine Learning Deep Learning Predicts the future Ex: Weather forecast Most AI used in practice is this type Expectations are high, but... Emergence Ex: ChatGPT, Stable Diffusion 5
OSS Mainstream AI development is not open source Some surrounding technologies are open source The Linux Foundation has compiled a Landscape, But it contains almost nothing related to Java AI software development is led by data scientists first As shown in the "AI & Data" summary, it's not targeted at application developers 6
AI The pace of AI technology evolution is fast Java's standardization and compatibility strategy, successful in the enterprise, does not align perfectly The success of AI-Infused Applications is key to leverage companies' Java assets and Java engineers. A balanced strategy of compatibility and innovation 8
Compatibility and Innovation Strategy Provide a framework to support building AI systems Validate with real-world use cases and systems that support vector DBs and GPUs Provide Java specifications and implementations that enterprises can confidently use, thanks to standardization Step 1 Practical Integration Step 2 Prototype Verification Step 3 AI Standardization 9
Requires the same enterprise-grade quality as existing systems: A software program with artificial intelligence capabilities Implements intelligent functionality by leveraging AI models Security Observability Safe Remote Call Transaction Resiliency Stable Data Access ・・・ 10
Meets AI Broadly speaking, AI-Infused App architecture is a distributed component system. This is precisely where Java, especially Jakarta EE, excels. Jakarta EE has a long track record of being used in enterprise systems where high reliability is required. Further, Jakarta EE standardizes specification in a vendor-neutral way, making it safe to use in enterprise systems. Jakarta EE is the best choice for building AI-Infused Application in enterprise systems 13
Control RAG Usage AI Framework Native Wrapper Training MCP / A2A / ACP Usage AI Cloud-Native Enablement AI Remote Access Vector API HAT Babylon Valhalla JCuda TensorFlow for Java FFM Deep Java Library Visual Recognition Jakarta Agentic AI MicroProfile REST Client LangChain4j Jakarta Messaging Quarkus AI Spring AI LangGraph4j Embabel Jakarta Transactions Jakarta Query Jakarta Data Jakarta Security Jakarta Restful Web Services Jakarta RPC MicroProfile Fault Tolerance MicroProfile Telemetry MicroProfile Config Jakarta Websocket Jakarta Config Agentic AI Tornado VM
toolkit for building AI-Infused Java applications Provides integration with numerous LLMs/SMLs Provides building blocks for common patterns (RAG, function calling) Abstractions for prompts, messages, memory, tokens, etc. Integration of diverse vector stores and document data 15
Enterprise-Grade Quality Jakarta Security MicroProfile JWT Authentication Jakarta Restful Web Services MicroProfile RestClient Jakarta Transactions Jakarta CDI MicroProfile Telemetry Authentication/Authorization Observability Safe Remote Call Transaction Resiliency MicroProfile Fault Tolerance Stable Data Access Jakarta Data Jakarta NoSQL 17 (in future?)
AI Defines common usage patterns and life cycles for AI agents running on Jakarta EE runtimes. The API will include a mechanism to define agent workflows. https://github.com/jakartaee/agentic-ai/blob/main/README.md#scope Defines integrations with other key Jakarta EE APIs such as Validation, REST, JSON Binding, Persistence, Data, Transactions, NoSQL, Concurrency, Security, Messaging, and so on. 26
key to Production-Ready AI-Infused Application is Enterprise-Grade Quality. Java and Jakarta EE are first citizens in this new era, building on their proven history in robust enterprise systems. While ensuring solid compatibility, Jakarta EE community actively embracing and advancing AI technologies for enterprise. 28