Technical debt increases costs, introduces security risks, and reduces business agility by hindering the adoption of new frameworks and cloud migration.
Legacy Java applications often contain outdated patterns like EJBs and critical vulnerabilities such as text4shell or log4shell that require identification and remediation.
Konveyor's rules engine uses the Language Server Protocol (LSP) to perform static code analysis, identifying specific incidents and migration paths in a codebase.
Konveyor AI combines static analysis with large language models using RAG and few-shot prompting to generate targeted code changes without requiring model fine-tuning.
A live demonstration shows how Kai automatically transforms a Java EE application to Quarkus by converting JavaX to Jakarta, JMS to reactive messaging, and EJBs to REST endpoints.
The tool focuses on migrating technology stacks, like from a Java EE monolith to a Quarkus monolith, but does not perform higher-level re-architecture or domain-driven design.
This approach provides structured, auditable migrations driven by static analysis rules, is model-agnostic, and generates a reviewable code patch for developers.
Developers interested in contributing to the project can join the Konveyor.io community through Slack, mailing lists, and regular community meetings.
26:55Shaaf
37:55Adam Bien
33:40Niklas Heidloff
39:10Niklas Heidloff
27:11Kevin Dubois & Daniel Oh
57:45Jonathan Vila
39:16Holly Cummins
44:40Alex Soto
Jobs that call for the skills explored in this talk.
Akros Ag
Akros Ag
Amdocs
Kontich, Belgium