Have you ever felt drawn in by the allure of AI as a Java developer, but found the terrain intimidating? While data science and machine learning can seem like uncharted territory, a bridge is now awaiting. Welcome to Spring AI, the Java application framework tailored for AI engineering. If you have been developing applications with the Spring Framework and Java, Spring AI is your seamless transition to the AI world.
Though Python reigns supreme in AI, the fact remains that many enterprise applications run on Java using the Spring framework. But here is the good news: with Spring AI, you will not feel lost in the AI maze. It offers a Spring-oriented API and abstractions, making AI application development a breeze.
Just imagine creating an AI app with a simple command.
spring boot new --name my-ai-app --from ai
This command, part of the new Spring CLI project, simplifies initiating AI projects and infusing AI into existing ones.
Harnessing the Spring ecosystem
Building an AI application is not just about connecting to the model. It necessitates robust extract, transform, and load (ETL) pipelines that bring your unique data to the AI model. ETL pipelines help enable use cases such as a Q&A over your documentation as well as a summarization of documentation. AI applications involve several interactions with one or more AI models. Using enterprise application integration (EAI) patterns, developers can set up conditional processing flows, even letting AI models dictate the direction.
Developers can use the array of projects offered by Spring, such as Spring Task, Spring Batch, Spring Integration, and Spring Data, in combination with Spring AI to provide the complete feature set of an AI application. Implementing emerging AI patterns, such as retrieval augmented generation (RAG), becomes a cinch with Spring AI.
Feature highlights
Spring AI provides portable access to diverse AI models, which, coupled with APIs, provides an AIClient abstraction that makes switching AI models easy.
A crucial part of AI applications is using a vector database. Spring supports multiple vector databases, and its portable API makes it easy to change implementations.
Crafting prompts for AI models can be intricate. The intuitive Prompt API provided by Spring AI, based on a template engine, makes this task straightforward.
Platforms
A key differentiator for VMware Tanzu is that the Spring team has worked closely with Microsoft to help developers use Spring AI with Azure’s OpenAI service. Developers can easily deploy Spring AI Applications to the VMware/Microsoft jointly engineered Azure Spring apps platform. Ask your VMware Tanzu or Microsoft representative for a free workshop or browse the workshop code and try it yourself!
The Spring team is also working with the VMware Tanzu Application Service team to provide the platform with an AI model tile that is backed by open source models such as those from Meta, and uses the latest PostgreSQL tile for vector database support.
Learn more
Spring AI is currently in an experimental and incubating status and aims to make AI accessible and intuitive for Spring Framework and Java developers. We are excited about its potential and plan to reach GA status by early next year. If you want to delve deeper, we have videos for those new to Spring AI, as well as a more advanced video, workshops, and more.
Dive in, experiment, and share your feedback on our GitHub project page.
