Who & why
Turns manual checks into a fast automated test suite that catches regressions in CI on every commit. Scales quality: what takes an hour by hand runs in minutes on every change. An engineer who writes code for quality (frameworks, e2e, test infra). Without one: manual regression each release (slow, gaps), prod regressions, no fast CI feedback, scary refactoring.
A day in the life
Morning: review the nightly CI run — real bug or flake; fix flakes. Day: write tests for new scenarios (e2e/integration/contract), grow the framework & fixtures, mind the pyramid level. Evening: integrate tests into CI/CD, improve speed & stability, expand critical-path coverage.
Key skills
Hard: programming, test frameworks (Playwright/Cypress/Selenium, pytest/Jest), API/contract tests, CI/CD, test data & environments, the testing pyramid, fighting flakiness, Page Object. Soft: engineering discipline (tests are code), systems thinking on coverage, pragmatism, dev communication.
Artifacts
Automated tests (e.g. Playwright), test framework, run reports, CI config. Automates scenarios from the Test case & Test plan; result is part of DoD (green CI).
How AI / vibe-coding boosts the role
Test from a scenario; flake fixing; coverage expansion; failed-run triage; Page Object refactor — with ready prompts.
Growth: Junior → Middle → Senior → Lead
Junior: tests by example in a ready framework. Middle: covers features, grows the framework, CI. Senior: test architecture & automation strategy. Lead/Automation Lead → QA Lead.
Common mistakes
Everything via e2e; tolerating flakes; hardcoded sleeps; brittle selectors; tests as unmaintained code.
What to learn
The testing pyramid, e2e/integration/contract tests, flakiness & determinism, CI/CD, Page Object, test data. Read: Playwright docs; xUnit Test Patterns; Fowler on test automation.
Salary (RU)
Junior ~90–150k₽/mo, Middle ~150–260k, Senior ~260–420k. Automation usually above manual; varies — check current data.
Laskoff agent mapping
No direct mapsTo; automation is core: a Playwright browser pass is a mandatory close step for web flows, with result verification by the quest-verifier agent. Tests are written by the engineer agent via TDD; Playwright MCP is the home e2e tool.
🤖 Persona prompt
You are an experienced SDET. Help me turn manual checks into a fast, stable automated suite in CI. Respect the testing pyramid: don't push everything into e2e. Write tests as clean maintainable code (Page Object, no duplication). Use role/aria/data-testid selectors, explicit state-based waits, no hardcoded sleeps. Fix flakes immediately — they kill trust. On a failed run, distinguish a real product bug from a test/environment issue. On request give Playwright/pytest tests, flake fixes and coverage expansion.