JVM Day: Three Tickets for Measurable AI Cases from the Java World (Category AI4SDLC)
I looked through the program for JVM Day 2026, which will take place on August 29 at T-Space in Moscow. I like that the conference is organized not around yet another list of new APIs, but around the things JVM engineers actually live with in production: performance, concurrency, migrations, correctness, security, and architectural trade-offs.
There will be 21 talks across three parallel tracks for people working with Java, Kotlin, and Scala. The program covers how JVM benchmarks mislead us, alternative concurrency models after Loom, Spring Data JDBC, Redis as part of a correctness model, JDK supply-chain security, and Mechanical Sympathy. I especially want to hear the talk about OpenRewrite: the T-Bank team says it automated 80% of a CI/CD migration.
There is also a direct bridge to AI. In a talk on formal verification in Scala, the speaker promises to show a multi-agent system that found 9 bugs, got 4 PRs into main, and opened more than 200 issues in the Scala ecosystem; according to the talk description, 82 of them have already been accepted and fixed by maintainers. This is exactly the level of discussion about AI that interests me: not “it became more convenient,” but a task, a method, a measurable result, and constraints.
I have three tickets to JVM Day, which will go to the authors of the three best stories about recent cases from the Java/JVM world where AI delivered measurable value.
Three formats are eligible:
- A case from your own practice—it may be anonymized;
- A link to a good public account of someone else’s implementation;
- An original research paper or research whitepaper.
In a comment on this post, briefly describe the chain: task → where and how AI was applied → baseline → result in numbers → how it was measured → how much verification and implementation cost. The metric may be lead time, migration or review duration, the number of defects and incidents, cost, performance, or the share of manual work. You may hide the company name and absolute figures, but the verifiable change must remain—for example, time fell by X%, while the share of work sent back fell by Y%.
An important constraint: the case must be recent. Your own result must have been achieved this summer and described for the first time in the submission; an external account, paper, or whitepaper must have been first published no earlier than June 1, 2026. A summer retelling of an old case will not qualify.
On August 21, I will write in the channel who gets the tickets and explain why. I should also note that a ticket covers admission to the conference, but not travel or accommodation.
JVM Day itself will take place on August 29 at T-Space, 7 Gruzinsky Val Street, Moscow. Registration starts at 09:30, talks at 10:40, and there will be an afterparty after the program. The detailed program and tickets are available on the conference website.
So bring your cases. I would like to assemble not another list of AI tools, but a small map of proven AI value in JVM development. We can discuss the cases in the comments, and meet in person at the conference, where I also plan to be.
#AI4SDLC #AI #Java #JVM #Engineering #Conference #Metrics