AI/ML Introduction for Java Developers

We're in a new era of software development driven by Machine Learning (ML), a subset of AI that predicts or generates results by analyzing large datasets and recognizing patterns. Combined with Java, ML can generate code, query unstructured enterprise documents, summarize emails, offer bug triage strategies, predict GC issues, and identify library inconsistencies.

This new phase offers exciting opportunities to boost productivity but also raises concerns for developers and enterprises. This session will cover AI/ML fundamentals, compare GenAI and PredAI, JSR #381 VisRec, prompt strategies, APIs, chatbot architecture, vector databases, and RAG strategies. We'll also explore REST-based APIs like Langchain4J, share advanced prompt techniques, and guide you on how to work with an LLM. Whether you're new or experienced with Java, you’ll find our code demos informative, highlighting where to apply GenAI effectively and where to avoid it.

Presented by Frank Greco (NYJavaSIG Chair) at JavaOne 2025 (CA, March 2025).

Tags: #Java #AI #ML #JavaOne

SES1126

[carlosabreu]

Albert Einstein reborn! I love this video! Thanks you so much

[sjzara]

There are some strangely critical comments. I really appreciated this introduction - I learned a lot.

[yadaom]

Awesome, I thought I will never understand how the LLM works. But after watching this video, my confidence it at the roof. I really liked the risk factor, how using AI can add risks and if the company at risk, then one should not be using the AI.

[VLADICAKG]

25:13 Legend! Translate this to Serbian! 🇷🇸 Pozdrav iz Srbije!

[MidnighterClub]

A very good talk. I found it very high level, then at the end Greco says "And you should implement one yourself to learn how it works." OK, but how? A talk on how to do just that, for beginners, would be appreciated I think.

[jchildress]

Great presentation, thank you for the overview and information.

[FlyingNeuron]

Finally a set of tools and apis for Java... Even Pytorch will be feeling Ashamed... Yeah!! I think Oh it's just another API Wrapper.

[johanmbaabu]

To me AI IS just software learning from datasets stored in its database . its just software like any other developed by computer engeeners just like me and you paid to maintain source code at a higher stake level

[chetAnrAo]

I thought they would show new coding practices but instead they just gave a road map for Java developers who want to learn AIML. But it would be fun if people started building ML and DL apps with Java.

[drownwarlord]

I can't even put a hope on this language anymore, I watched this with a good assumption this will introduced me a new Java API for AI/ML, and yep, turned out just another HTTP wrapper slop. Get well soon.

[ArkenGAMES]

I wish there was more talk about self-hosted models and fine tuning.

[skyheart]

Duke does not have visible eyes! please understand that.

[metalToad]

This is not determinizedm but actually every AI system is it, thank good and i hope AI could'nt think about any real things anless u could call it Skynet

[rural_world]

That's how it sound when some one knows about AI.

[tibordigana]

First of all 1. I need to have people around me, and not the machines. Second, 23 years ago we had courses at the Helsinki University of Technology, my teacher was Prof. Heikki Koivo, we did not call it AI because it was Neural Nets, Fuzzy logic and Genetic algorithms. Start thinking what you really need!!!