The eBook is available via O’Reilly and Amazon, and seeing it there still feels slightly unreal. Writing a book is not one big heroic effort. It is many small, unglamorous moments stitched together. Early mornings before work. Evenings when everyone else is already asleep. Notes written on trains, in hotel rooms, or between meetings. Life does not pause just because you decided to write a book.
This one was written alongside family life, a full-time job, travel, and a technology landscape that keeps shifting under your feet. That constant movement shaped the book more than any initial outline ever could.
What I wanted to avoid from the beginning was writing yet another abstract AI book. There are plenty of those already. Interesting, yes, but often disconnected from the reality of enterprise software. In real projects, AI is not a playground. It has to fit into existing systems, comply with constraints, behave predictably, and not explode operational costs.
That is why the word applied matters so much to me.
This book is built from what we actually see in the field. From conversations with teams who are already deploying AI into production systems. From architectures that survived first contact with reality. From mistakes that hurt enough to be remembered. Together with Alex Soto Bueno and Natale Vinto, we tried to extract patterns that Java developers can realistically apply without throwing away everything they already know.
The focus is on how generative AI, LLMs, and machine learning integrate into the Java enterprise world. Not as a side experiment, but as part of real systems. We talk about architecture, integration patterns, inference APIs, and the practical implications of running AI workloads in production.
Frameworks like Quarkus play an important role here. Quarkus has quietly become a very strong foundation for AI-infused Java applications. It is fast to iterate on, efficient at runtime, and fits naturally into modern deployment models. Combined with projects like LangChain4j, it allows Java developers to explore AI without leaving their ecosystem behind.
If you have been reading my recent articles, you will notice familiar ideas. Topics like local models, agent-style workflows, memory management, and production constraints show up both in the book and in my writing elsewhere. They all come from the same question: how do we keep Java relevant and useful in a world where AI is becoming part of everyday software?
All code examples from the book live in a public GitHub repository: https://github.com/applied-aI-for-enterprise-java-book
That repository is intentionally treated as a living companion. The ecosystem moves fast, and code needs to move with it.
Around the same time as the book, I also published a three-part series on O’Reilly Radar called The Java Developer’s Dilemma. Those articles reflect on the same tension many of us feel. How do we stay grounded in our craft while the tools, expectations, and problem spaces evolve so quickly? I do not believe Java is falling behind. I believe it is adapting, often more quietly than other ecosystems, but no less effectively.
This book exists because of the people building critical systems today and asking hard questions about AI, reliability, and responsibility. It is a snapshot of what we have learned so far, knowing full well that the story will continue to evolve.
If you read the book and find it useful, a short review on Amazon or O’Reilly goes a long way. Feedback matters, especially for technical books that try to stay practical instead of flashy.
And yes, it probably also works surprisingly well as a Christmas present for a Java developer. At the very least, it pairs nicely with coffee and a slightly skeptical mindset.
Markus
