Work
Each of these is a real engagement or product. Problem, approach, and the numbers that actually moved.
Turns a URL into a reviewed, risk-rated test suite in under 10 minutes.
An AI agent that crawls a web app and writes the test cases — then turns them into Playwright or API scripts.
- 2 days → 9 min
- Time to first test plan
- 78%
- Cases surviving human review
- +31%
- Flows found vs. manual audit
Next.jsPlaywrightClaude Agent SDKPostgresDrizzle
Cut the nightly regression run from 6 hours to 11 minutes and flake to under 1%.
Inherited a 2,400-test suite with a 40% flake rate that nobody trusted. Rebuilt it into a gate the team actually respected.
- 6h → 11m
- Suite runtime
- 40% → 0.8%
- Flake rate
- -64%
- Escaped defects / quarter
PlaywrightTypeScriptGitHub ActionsDocker