Stop Prompting. Start Building Software Factories for Java with Ingo Eichhorst

AI can write code. Most developers have seen that. The harder question is how to make AI produce good software consistently. In this live coding session, I am joined by ‪@ingo-eichhorst‬ to explore the idea of Software Factories for Java development. Instead of asking an AI to implement a feature and hoping for the best, we will start by defining what a good result looks like. We will establish clear quality criteria, create ways to evaluate the outcome, and then let AI do the implementation work. Using a real open source Java project, we will add a new feature while exploring a different approach to AI assisted software development. The goal is not to build an autonomous developer. The goal is to build a repeatable process that produces useful results. What you will see We will work on a real Java codebase and: Select and define a new feature Describe the desired outcome Define measurable quality criteria Use AI to implement the feature Evaluate the result against our criteria Improve the process when the results are not good enough Topics we will explore Software Factories AI assisted development Feature implementation with AI Evaluation driven development Claude Code and AI agents Defining quality in software engineering Human oversight and feedback loops Practical Java development Why this matters Many teams focus on prompting AI more effectively. Software Factories take a different approach. They focus on defining quality, measuring outcomes, and automating repeatable processes. This shift turns AI from a coding assistant into part of an engineering system. Join the conversation As always, audience participation is encouraged. Challenge our assumptions, question the results, suggest better evaluation criteria, and help us decide what "good" really means. Who this is for Java developers Software architects Engineering leaders AI practitioners Developers interested in practical AI workflows By the end of the session, you will have seen how a Software Factory approach can be applied to a real Java project and why defining success may be more important than writing prompts.