Projects
Tools I've built for quality
QA Intelligence
CI Intelligence Engine for Playwright
A production-ready CI intelligence engine installable into any Playwright project. Provides PR failure diff analysis, retry-aware flaky detection, recurrence tracking, and automated GitHub PR comments, blocking only on new non-flaky regressions.
Tech stack
Challenges
- →Designing retry-aware failure classification (New / Flaky / Still Failing / Fixed)
- →Building CLIs that integrate into existing CI pipelines without framework lock-in
- →Balancing PR blocking strictness with flaky test tolerance
Lessons learned
- ✓CI intelligence is as valuable as test coverage. Teams need context, not just pass/fail
- ✓Packaging as npm module dramatically lowers adoption friction vs. monolithic frameworks
- ✓Baseline comparison transforms noisy CI into actionable release signals
QA Copilot
AI-Powered QA Assistant
An AI-assisted platform that transforms product requirements into comprehensive QA strategies, analyzing business rules, risks, missing information, automation candidates, and generating Playwright test skeletons with secondary LLM quality evaluation.
Tech stack
Challenges
- →Structuring multi-step LLM workflows for consistent, actionable QA deliverables
- →Supporting multimodal input (screenshots, recordings) across providers
- →Building secondary evaluation pass to measure coverage and identify gaps
Lessons learned
- ✓AI augments QA thinking. It doesn't replace risk assessment and domain judgment
- ✓Structured prompts with type-specific strategies produce significantly better output
- ✓Quality evaluation of AI-generated tests is essential before trusting automation candidates
Playwright Framework
Scalable Automation Architecture
A production-grade Playwright automation framework demonstrating scalable test architecture with Page Object Model, test tagging, structured logging, Docker support, CI integration, and rich artifacts: screenshots, video, and traces.
Tech stack
Challenges
- →Designing modular fixtures and hooks for reusable test setup across suites
- →Balancing test isolation with efficient parallel execution in CI
- →Integrating artifact capture without bloating pipeline runtime
Lessons learned
- ✓Framework structure matters more than test count for long-term maintainability
- ✓Tagging and selective execution are critical for fast feedback loops
- ✓Docker reproducibility eliminates 'works on my machine' CI failures
QA Challenge
A new checkout flow is ready for production. Would you approve the release? Investigate the flow, log findings, and reveal a professional QA release review.
Investigate
Review a checkout release candidate: cart, promo codes, shipping, payment, and order payloads.
Log findings
Document risks in your own words. Smart matching recognizes related issues without exact titles.
Release review
Reveal a senior QA review covering coverage, critical findings, release risk, and retest approach.