Join us for a great update with Markus Eisele (https://www.linkedin.com/in/markuseisele/) about IBM Bob and Java topics. 

An AI coding agent can produce a convincing diff. That does not mean the application builds, the tests pass, or the developer saved any time. For Java developers, the real question is not how much code IBM Bob can generate, but how to work with it so that changes remain understandable, reviewable, and verifiable.
In this session, Markus shares the patterns that make Bob more effective on real Java projects. We’ll look at establishing a clean baseline, providing the right project and build context, protecting files that should not be touched, and breaking larger tasks into small, verifiable steps. A live demo will follow a Java task from analysis through code and dependency changes to build validation, tests, and review.
We’ll also examine the less polished cases: what to do when the agent stalls, repeats work, changes too much, or leaves manual cleanup behind.
Finally, we’ll discuss how to judge the result. Not by prompts, generated lines, or Bobcoins alone, but by completed work, passing builds, developer confidence, manual intervention, and actual time saved.

#events-home