Agents fail in ways ordinary software does not. We need people who enjoy finding exactly how, and who write it down so it never happens twice.
What we’re looking for
- Manual and automated testing. You should be fluent in both, not one
- Playwright for end-to-end automation; the Playwright MCP is a strong advantage
- Writing test plans, test cases and regression suites that outlive the sprint
- API testing with Postman or equivalent, including auth and error paths
- Cross-browser and real-device testing on mobile
- Bug reporting discipline: reproduction steps, environment, expected versus actual
- Running tests in CI/CD so failures block a release rather than surprise it
- Basic TypeScript or JavaScript, enough to read the code you are testing
- Testing non-deterministic AI output: prompt regressions, tool-call failures, hallucination checks
- Genuine attention to detail, and the patience to reproduce the hard ones
Break something of ours. Find a bug on this site or any product we have shipped, report it properly, and attach it to your application.