Contract & overflow QA · European hours
I assure software that agents wrote.
Your team ships code faster than anyone can review it. I build the regression suites, the defect evidence and the release confidence that make that safe — and I write the specification the output gets held to before the agent runs.
Two decades assuring software in banking, healthcare and licensed multi-jurisdiction online gaming — industries where an outcome has to be traceable, documented, defensible and repeatable. Based in South Africa, working European hours, with a full working-day overlap with Ireland and the Netherlands.
- Checks run
- 80
- Failures
- 0
- Deployed
- 1 min ago
Every deploy is gated by this site's own regression suite, an axe-core accessibility scan and Lighthouse budgets. Nothing ships red. View the run →
What I do
- 01
Assurance for AI-generated code
Acceptance criteria written before the agent runs, and evidence the output matches them afterwards. Scope, deliverables, observable criteria and dependencies specified up front, then the result reviewed against that specification rather than accepted on trust. This is the thing most teams using agents have no process for yet.
- 02
Test automation from zero
Your app has no automated tests, or a suite nobody trusts. I build a Playwright regression suite around your critical user journeys and wire it into your CI, so every deploy is checked before your customers check it for you.
- 03
Overflow QA for agencies
White-label senior QA capacity for delivery crunches: test planning, exploratory testing, and structured defect reports with evidence your developers can act on immediately. NDA-ready, your branding, your process.
- 04
Release readiness and defect evidence
Independent pre-release testing with reproducible, annotated defect reports — screen capture, steps, environment, severity — so nothing ships on a hunch and nothing bounces back as "cannot reproduce".
Playwright · pytest · GitHub Actions · exploratory & structured manual testing · spec-first documentation · decision records
Start small: a fixed-price pilot sprint.
One sprint. I take one project or one critical user journey, build the automated test coverage for it, and hand over the suite, the CI integration and a findings report. Fixed price agreed up front — you see exactly what working with me looks like before committing to anything bigger.
€1,950 · two-week pilot
Ongoing engagements from €320/day · €1,500/week
How I got here
I spent my career in places where testing is not optional: Home Loans at Standard Bank through the National Credit Act rollout, healthcare systems at Hospiserve as test analyst and then test lead, and eight years as a technical tester and QA engineer at Digioutsource, inside licensed multi-jurisdiction online gaming. I owned test strategy, and the documentation I wrote was adopted as the company's source of truth.
Over the past year I have been proving those practices out against a new reality — software written by AI agents. Every project on the Work page was specified by me first and implemented by an agent, then reviewed against that specification. The recurring artefacts are the point: phase decision records, ticketed backlogs with dependency graphs, and test plans whose case IDs stay stable across unrelated codebases.
For larger engagements I work with a small network of experienced testers I have personally worked with — the site is me, the capacity can scale.