Believe · Product & developer tooling
Build the product.
Equip the team.
Two complementary parts of my work: delivering an internal application and building tools that help teams automate interfaces.
01 / beOne
From needs
to delivery.
MVP · Delivered to integrationAn internal product
beOne brought together product framing, UX/UI and implementation. My contribution covered the MVP, interface development in React and TypeScript, and the delivery workflow.
The deployed MVP gave the team an application to use and evaluate as the product took shape.
My contribution
- Product scope and user experience for the MVP.
- React and TypeScript application development.
- Contribution checks, versioning and CI configuration.
- Delivery to integration through the GitLab, Argo CD and GCP workflow.
A delivery path with explicit stages
- ChangeDevelop and review the application.
- BuildPackage a version through CI.
- DeployUpdate the integration environment through GitOps.
- EvaluateUse the deployed MVP to assess the product experience.
02 / INTERFACE MAPPER
Turn a page
into an automation map.
Interface automation repeatedly needs the same groundwork: understand the page, find its elements and choose selectors. Interface Mapper centralises this work behind an MCP server using Playwright.
How it works
- Inspect a web page through Playwright.
- Extract elements and candidate selectors.
- Prefer test attributes or identifiers where available.
- Check selector uniqueness and maintain the page map.
- Use the map to support Cypress command generation.
Why it matters
A page map makes the interface structure inspectable and reduces repeated selector work. MCP exposes these capabilities as tools that an assistant can call.
Selector checks establish uniqueness; business scenarios define which behaviour to verify. My work connects this tooling with the shared test platform.
Browser add-ons
I also developed a browser add-on that analyses the DOM to generate Cypress tests and commands. It complements the reusable test components by helping authors prepare the automation around a page.
A separate Faker extension prepares test data. Each tool addresses a specific part of authoring and investigating a test.