AI Output Gets Better When Your Workflow Gets Stricter The biggest gains in AI-assisted Java development often come from changing the environment around the model: better context, tighter review loops, and earlier security boundaries.

Better AI output often comes from a stricter workflow, not smarter prompting.

Context quality, instruction files, validation loops, and security hygiene matter more than people think.

bazlur.substack.com/p/ai-output-...

I'll be speaking at Microsoft JDConf 2026 on Wednesday, April 8, at 2:00 PM EDT

Talk: Building LLM Apps in Java with LangChain4j

This talk is for Java developers who want to integrate LLMs into real backend systems, not toys.

Free and virtual. Register: jdconf.com

Before You Ask AI to Code, Write a Better Spec In AI-assisted development, ambiguity does not go away. It turns into code that looks plausible. One of the fastest ways to improve output is to refine what exists before implementation begins.

AI does not remove ambiguity.
It amplifies it.

If the spec is weak, the implementation will often be fluent and wrong in ways that are harder to catch later.

Better contracts matter more now, not less.

bazlur.substack.com/p/before-you...

Code Is Cheap. Trust Is Expensive The real bottleneck in AI-assisted Java development is no longer producing code. It is understanding it, verifying it, and deciding whether it deserves to be trusted.

AI has made coding cheaper. It has not made judgment cheaper. That’s the main idea behind the first article in my new series for Java developers:

read it: open.substack.com/pub/bazlur/p...

Excited to share that I’ve contributed two chapters to Sustainability for Java Developers, a collaborative effort by Java Champions and community experts.

If you care about building better Java systems with a lighter environmental footprint, this one’s for you 🌱☕

Modern Java deserves modern concurrency.

If you haven’t grabbed your copy of Modern Concurrency in Java yet, now’s a good time.

bit.ly/3KEhqCl

Enterprise AI Conference for Developers & Architects | QCon AI NYC 2025 QCon AI New York (Dec 16-17): The AI conference for senior engineers & leaders scaling reliable, production-grade AI. Get real-world patterns. Register!

Excited to be joining the QCon Certification Program to deepen my skills and stay current with industry practices.

Looking forward to learning from experts and applying it in real projects.
certification.qconferences.com

#ContinuousLearning #SoftwareArchitecture #QCon

The JUnit team has released JUnit 6.0.0 (GA on September 30, 2025), unifying version numbers across Platform, Jupiter, and Vintage and raising the minimum to Java 17 (and Kotlin 2.2 for Kotlin test code).

Read more 👉 https://lttr.ai/AlIPI

#Testing #Java #Kotlin

Kotlin users can simplify coroutine testing with direct suspend methods, and JFR integration becomes easier to adopt for performance investigations.

Read more 👉 https://lttr.ai/Ak1yZ

#Testing #Java #Kotlin

For most teams already on Java 17 and JUnit 5.14, adoption should be a routine dependency bump followed by a quick dry run, modernizing any build plugins (e.g., Surefire/Failsafe ≥ 3.0), and validating CSV-driven tests.

Read more 👉 https://lttr.ai/AkqJz

#Testing #Java #Kotlin

JUnit 6.0.0 Ships with Java 17 Baseline, Cancellation API, and Kotlin suspend Support: https://lttr.ai/AkkEC

#Testing #Java #Kotlin

Tons of comments in the code

JUnit 6.0.0 Ships with Java 17 Baseline, Cancellation API, and Kotlin suspend Support Introducing JUnit 6.0.0: a transformative update that unifies versions, elevates minimum requirements to Java 17, and introduces streamlined support for Kotlin suspend tests. Enjoy enhanced testing pe...

Introducing JUnit 6.0.0: a transformative update that unifies versions, elevates minimum requirements to Java 17, and introduces streamlined support for Kotlin suspend tests. Enjoy enhanced testing performance with the new CancellationToken API, built-in JFR listeners and many more—migrate today!

Fresh off the press. Alhamdulillah!

My book, 'Modern Concurrency in Java,' is finally available.

learning.oreilly.com/library/view...

Amazon: www.amazon.ca/Modern-Concu...

Thanks, it's fixed now.

I'm also wondering that!

Modern Concurrency in Java Welcome to the future of Java. With this book, you'll explore the transformative world of Java 21's key feature: virtual threads. Remember struggling with the cost of thread... - Selection from Modern...

I'm happy to share that the Early Release of Modern Concurrency in Java, with all chapters, is now available on the O'Reilly platform.

If you have access to the O'Reilly platform, you can take a look now.

learning.oreilly.com/library/view...

Building Smart Mock: An AI-Powered API Mock Server with Spring Boot and Ollama In software development, teams often need to mock APIs during the early stages of integration.

Over the weekend, I built Smart Mock—an AI-powered API mock server that generates realistic, context-aware responses from your OpenAPI specs.

🛠 Built with Spring Boot, Ollama, and LangChain4j
💻 Runs locally—ready in minutes

Check it out here: bazlur.substack.com/p/building-s...

Fail-Fast vs Fail-Safe Iterators in Java: Understanding ConcurrentModificationException I am pretty sure you have encountered a ConcurrentModificationException when attempting to modify a collection within a loop. Once you encounter it once, you learn the lesson that this is something we...

Let's revisit the very familiar ConcurrentModificationException and why it happens. But more importantly, why does Java NEED to throw it?

www.linkedin.com/pulse/fail-f...

I've launched The Coding Café on Substack: lnkd.in/gtkcVUR2

The subscription is growing rapidly!

☕ What to expect:
- Deep dives into Java features
- LLMs and AI with Java experiments (LangChain4j and beyond!)
- Trendy stuff that matters

Join me to get LLMs and Java in your mailbox weekly!

Building FormPilot: My Journey Creating an AI-Powered Form Filler with RAG, LangChain4j, and Ollama Have you ever found yourself filling out the same information on web forms over and over again?

Built an AI Chrome extension that fills forms intelligently! 🚀

FormPilot uses: - Local LLMs (privacy-first)
- RAG for personalization
- Java Spring Boot backend
- Zero API costs.

Code + guide:

open.substack.com/pub/bazlur/p...

Building an AI-Powered Flight Tracker with Spring Boot and Local LLMs Build an AI-powered flight tracker with Spring Boot and LangChain4j. Uses local AI models, real-time aircraft data, and runs entirely offline. Complete tutorial with code.

🛩️ I built an AI that knows what planes are flying overhead — in real time.q

No cloud. No APIs. No fees.
Runs fully local.

Built with:
⚙️ Spring Boot + LangChain4j
🧠 Ollama (local AI)
📡 Real-time flight data

What would you build with local AI?

bazlur.substack.com/p/building-a...

Read the rest of the interview:

www.infoq.com/news/2025/06...

Introducing Embabel: Advanced AI Agent Development for Java Applications Introducing the Embabel Agent Framework, a pioneering platform developed by Spring founder Rod Johnson, designed to revolutionize AI applications on the JVM. By integrating structured agent developmen...

www.infoq.com/news/2025/06...

🚀 Spring founder Rod Johnson just launched Embabel, and Java developers can finally build AI agents without Python!

This isn't just another wrapper. And he's actually pulling it off.
Give it a read.

Introducing the Embabel Agent Framework, a pioneering platform developed by Spring founder Rod Johnson, designed to revolutionize AI applications on the JVM.

Read more: www.infoq.com/news/2025/06...

Building Robust AI Applications with LangChain4j Guardrails and Spring Boot - A N M Bazlur Rahman Learn how to build secure AI applications using LangChain4j guardrails in Spring Boot. Implement input/output validation, prevent prompt injection & more.

Guardrails are finally available in LangChain4j! I explored them this morning—super easy to use and a big step toward safer AI apps in Java.

Check out my write-up:

bazlur.ca/2025/06/21/b...

Introducing the Embabel Agent Framework, a pioneering platform developed by Spring founder Rod Johnson, designed to revolutionize AI applications on the JVM.

Read more: www.infoq.com/news/2025/06...

Introducing Embabel: Advanced AI Agent Development for Java Applications Introducing the Embabel Agent Framework, a pioneering platform developed by Spring founder Rod Johnson, designed to revolutionize AI applications on the JVM. By integrating structured agent developmen...

🚀 My latest news piece is live on InfoQ: www.infoq.com/news/2025/06...

I had the pleasure of speaking with Rod Johnson, the creator of Spring and Embabel, to gain insights into this framework for building AI-powered Java applications.

Check it out and let me know your thoughts!

Java turns 30 today! 🎉

From "Write Once, Run Anywhere" to powering cloud-native, AI-driven, and distributed systems, Java has grown into a modern, expressive, and resilient language—far beyond the stereotypes.

Happy birthday 🎂 to Java

Instance Main Methods Move from Preview to Final in JDK 25 JEP 512 enhances Java with Compact Source Files and Instance Main Methods in JDK 25, easing the learning curve for beginners. Key features include simplified syntax, automatic imports, and a new IO he...

www.infoq.com/news/2025/05...

JEP 512 enhances Java with Compact Source Files and Instance Main Methods in JDK 25, easing the learning curve for beginners. Key features include simplified syntax, automatic imports, and a new IO helper class, to create programs with minimal complexity.