I build practical test automation and use AI where it improves speed, clarity, and feedback — with human review, deterministic checks, and visible guardrails.
| Project | What it shows | Stack |
|---|---|---|
| Playwright AI Storefront | UI automation, Page Objects, API mocking, reports, Copilot workflow | Playwright, TypeScript, GitHub Actions |
| API Testing + AI Insights | Contract tests, negative coverage, idempotency, AI failure triage | Playwright API, Zod, Express, GitHub Models |
| AI Test Case Generator | Structured test design from user stories with schema guardrails | TypeScript, GitHub Models, Zod |
| Cursor + MCP QA Agent | A real MCP server, agent rules, reusable QA workflows | Cursor Agent, MCP, TypeScript |
- Start from product risk and observable behavior.
- Keep automation readable, reliable, and useful in CI.
- Use AI for drafts, exploration, and triage—not as a source of truth.
- Validate model output and keep assumptions explicit.
- Treat traces, contracts, and reproducible evidence as part of the deliverable.
Playwright · TypeScript · API testing · GitHub Actions · GitHub Copilot · GitHub Models · Cursor · MCP
Each featured repository includes a quick demo, architecture notes, automated checks, and an interview walkthrough.