The frontend left the monolith, and a team owns it.

The frontend was part of a Java monolith. It was built and released with that application, so every frontend change moved at the monolith's pace and through the monolith's pipeline.

The team that owns that frontend now did not exist at the start of the contract. Hiring it was part of the job, which meant the work was never only a migration — the people who would maintain the result had to be there before the result was finished.

I helped hire the three engineers who own the frontend now, and I trained them on it.

Training is where a migration lands or quietly fails. A standalone app with its own pipeline is only an improvement if the people inheriting it can change it without me, so the target I worked toward was not a finished migration — it was three engineers who could take the next decision themselves. The contract ended in December 2023 with the codebase in their hands.

The frontend came out of the monolith as a standalone React and TypeScript application with its own CI/CD. The pipeline is what makes a split like that real: until the frontend can be built, tested and deployed on its own, moving the code has only changed where it lives.

At the end of the contract, in December 2023, the frontend was a standalone React and TypeScript app with its own CI/CD pipeline, and three engineers owned it.

I would rather be judged on the second half of that sentence than the first. A migration that finishes with one person holding all the context has moved the risk rather than removed it. This one finished with a team who could take it from there.