Context
My Wolt team ran weekly regression testing alongside feature development. We also held manual-testing sessions, which required test cases, Confluence pages, and Google Sheets for the team.
I worked on both sides of that process by maintaining UI tests and automating preparation for the manual sessions.
UI regression tests
I added and maintained Maestro UI tests for the team's weekly regression testing.
Manual-test preparation
I wrote Python scripts and agent-assisted workflows to prepare team manual-testing sessions. The workflow drafted test cases for review, then generated the Confluence pages and Google Sheets used to coordinate the session. It removed multiple hours of repetitive setup and coordination each time.
Feature validation
For features I owned, I checked accessibility and localization during implementation and regression testing, making sure each feature worked across the app's languages and locales, from Japanese to right-to-left Arabic and Hebrew.
Telemetry Viewer also contributed to release validation. The iOS app streamed analytics events over a local WebSocket to a macOS client with search and filtering. I used it to verify what a feature emitted, investigate unexpected events, and check experiment instrumentation without covering the phone with the old full-screen event viewer.
What I'd do next
If I returned to this work, I would document which regression checks belonged in Maestro, which required a manual session, and which needed telemetry verification. I would also measure setup time and recurring failures before choosing the next task to automate.