Dwane Debono · Principal QA

I help teams build quality in, not test it in at the end.

I'm Principal QA at Find.co, working through my Estonian company, Debono Consultancy OÜ. I shape test strategy with product and engineering, build automation that gives fast, honest feedback, and pull performance and security checks earlier, where they're cheap to fix.

Portrait of Dwane Debono

Featured project

A testing lab you can click through

Max Profit Calculator

A small stock-picking calculator (Spring Boot API and React UI) that I use as the system under test for nearly every testing technique I care about. The site walks through the test pyramid, the techniques behind each suite, and how to test an AI agent, with live reports from CI.

The calculator's API runs on a free tier, so the first calculation can take around 30 seconds while it wakes up.

  • Each piece of logic is checked on its own, in milliseconds. Property-based tests go further: jqwik generates hundreds of random inputs to prove a rule holds for all of them, not just the examples someone thought of.
  • Pitest plants small bugs in the code, like flipping a < to <=, and reruns the tests. A bug that survives means a missing test. The build fails if fewer than 90% are caught.
  • The UI records exactly what it expects from the API, and the API proves it delivers that. If either side breaks the agreement, the build fails before anything is deployed.
  • Scenarios written in plain Given, When, Then language drive a real browser. The specs read like requirements and double as living documentation.
  • The real Docker images start exactly as they would be deployed and are tested over the network. Stress tests then push the load up to find where performance gives way.
  • Schemathesis reads the API description and fires hundreds of generated requests, odd and invalid ones included. Every response must match what the API promises.
  • The AI advisor's answers are replayed from recordings and scored by a second model acting as judge, so answer quality is measured rather than eyeballed.

How I work

Four ideas I keep coming back to

Confidence

Coverage isn't confidence

A green build only means something if the tests would fail when the code breaks. Mutation testing is how I find out.

What I'm building

Work, writing and the odd side project

Off the clock

A smart home, built like software

Most of my tinkering goes into my home, which runs on Home Assistant. I keep it local-first and private, and treat automations like any other code with real users: clear intent, sensible fallbacks, and a way to notice when something quietly stops working.

About Home Assistant

Writing

TestAutonation

I edit TestAutonation on Medium and write hands-on guides for testers, from WebdriverIO and Gatling to OWASP ZAP and mutation testing.

Read on Medium

Open source

Testing playgrounds

Smaller public repos where I try out ideas: an API test framework, Cucumber results flowing into ReportPortal, and visual testing with Applitools.

See them on GitHub

Toolbox

What I reach for

Practice
Test strategy · Risk-based testing · Shift-left · Quality coaching
Automation
Playwright · WebdriverIO · Cucumber · API testing · CI/CD quality gates
Contracts and APIs
Pact · OpenAPI · Schemathesis · Testcontainers
Performance
Gatling · Timings API · Browser Performance API
Security
OWASP ZAP
Test effectiveness
Mutation testing · Pitest · Property-based testing
AI agents
Agent evals · Recorded replays · LLM-as-judge