SDET Roadmap 2026: 90 Days from Scripts to Quality Gates

Pick login + one money path. Rewrite them with role locators. Put them on a merge gate. If you are still on Selenium, run the hybrid cutover in Playwright vs Selenium instead of translating the whole pack.

• Week 1: install Playwright, one spec, trace on failure.
• Week 2: move that spec into CI with npx playwright install --with-deps.
• Week 3: tag @critical and make it the only required check.
• Week 4: teach the team to open traces instead of rerunning blindly.

npm init -y && npm i -D @playwright/test
npx playwright install --with-deps chromium
npx playwright test tests/login.spec.ts --trace on

Install the flake-aware gate from CI/CD quality gates. You are done with this month when a settings flake cannot block a refund fix, and a refund break cannot merge.

• Log first-fail-then-pass as flake, not as success.
• Cap retries at 1 on @critical.
• Name an owner for each quarantined test.

Add a supervised agent loop from the Playwright test agents setup. The agent may draft specs under tests/agent/. You merge them. Then add one LLM or RAG fixture if your product has a model — see the LLM evaluation framework.

## Weekly deliverable (days 61–90)
- Mon: agent drafts one spec from a written AC
- Tue: you reject or rewrite assertions
- Wed: CI runs the merged file
- Thu: one eval fixture or one mutation on an agent PR
- Fri: note what the agent was not allowed to touch

If your org needs this as a factory rather than a personal 90-day sprint, use AI quality engineering services.

• Pass (day 90): A broken refund cannot merge; an agent-only 200-assert cannot merge; one eval fixture can fail CI.
• Fail: You collected courses and still cannot point at a required GitHub check.
• Cleanup: Delete personal "sandbox" specs that never ran in CI. The roadmap is the gated files, not the bookmarks.