Harness Engineer (SDET / QA Automation)

At Strattum, most of our code is born from AI agents. That changes everything about quality: code ships fast and in volume, and the bottleneck is no longer "writing the feature" — it's proving, automatically and reliably, that what the AI generated is correct, before it reaches production.

That's where you come in. We're looking for someone to build and maintain our test harness: the verification system that wraps our entire development process. This is not manual QA. It's test infrastructure engineering — fixtures, ephemeral environments, runners, CI orchestration, reports, and the feedback loop back to the agents — that lets us ship fast without breaking things.

If treating the test suite as an engineering product doesn't excite you, this role isn't for you.

What you'll do:

• Build and maintain the harness end to end — fixtures, factories, data seeding, mocks/stubs, ephemeral environments (Docker Compose / throwaway containers) and runners that run locally and in CI without flakiness
• Deliver E2E tests for back-end and front-end covering real user flows and contracts between services
• Design automated verification for AI-generated code — executable guardrails that validate every PR (e2e, contract/API, regression, smoke) and return actionable feedback to the agent loop
• Own the quality pipeline in CI — GitHub Actions: parallelization, caching, merge gates, flaky-test quarantine, feedback time
• Attack flakiness and runtime as product metrics — a slow or unstable suite is your bug
• Define the testability standard the other engineers (and the agents) follow

On seniority: we don't measure by years on paper. We measure by autonomy: if you deliver end-to-end E2E tests for back-end and front-end on your own, from strategy design through to a stable green in CI, you have the seniority we're after.

Required:

• Real experience building test automation as infrastructure (SDET, Test Infra, senior QA Automation) — not just writing one-off cases
• Command of at least one E2E stack (Playwright, Cypress, Selenium or equivalent)
• Comfortable across back-end and front-end — full-stack from a testing lens
• Fluency in CI/CD and the fight against flakiness

Nice to have:

• Testing LLM/agent-based systems — evals, non-deterministic tests, contract testing of AI outputs
• Visual regression (Playwright snapshots, Chromatic, Percy)
• Contract/API testing (Pact, schemathesis) and data-pipeline testing
• Having used Claude Code or AI codegen tools day to day

How to apply: no résumé — we want to see you doing it. Record a video of up to 5 min showing an automated test environment you built, explaining what you did and why (something you already created, or something built for this application). Show what you built, the decisions you made, and how you handled the hard parts (flakiness, state isolation, CI). A screen recording with your voice is enough — clarity of reasoning matters more than editing. Paste the video link in the application form.