Refounding a European Telecom Self-Service Portal

Context

An European telecom operator relied on a customer self-care portal built in React. After sustained delivery problems and a growing bug backlog, the client escalated and the project needed a credible path back to stable releases.

Problem

Large parts of the frontend were difficult to extend safely. Bugs accumulated faster than the team could resolve them, new features were slow to ship, and confidence in the codebase - and in the delivery relationship - had eroded.

What I owned

I refactored a major portion of the web application to eliminate outstanding bugs and introduce a new app architecture with a maintainable React codebase. I developed new features on top of that foundation, optimized the build and deployment process, and took a leadership role coordinating frontend delivery.

I acted as the bridge between frontend and backend teams, mentored frontend colleagues, and helped define patterns that made subsequent feature work significantly faster.

Interesting engineering parts

The work was less about a single clever feature and more about systematic recovery: isolating unstable areas, rebuilding them with clear state management and component boundaries, and proving the new architecture could scale to ongoing feature delivery without reintroducing the same class of bugs.

The refactor had to happen while the product remained in active use - so changes needed to be incremental enough to ship, but structural enough to actually fix the underlying problems.

Outcome

The bug backlog dropped from over a hundred issues to a handful. Delivery speed improved, reliability increased, and the team regained client trust. New features shipped on the refactored foundation without the same instability patterns returning.