AutoQA

In an ecosystem of multiple development cells with hundreds of weekly deployments, manual QA became a critical failure point. The organization was losing dozens of engineering hours daily to visual inspections prone to human error, hindering department scaling and risking the integrity of analytics and WCAG accessibility.

I implemented a hybrid architecture based on Electron and Next.js 16 to decentralize processing. By distributing the computational load to each developer’s local machine, we eliminated the need for expensive central servers and optimized Puppeteer’s scraping engine performance. The workflow was finalized with a GitHub Actions CI/CD pipeline to ensure environment parity across Windows and Linux.

AutoQA compressed entire QA workdays into processes of minutes, freeing up team bandwidth for product development. We achieved 100% DOM and sitemap coverage on massive sites (>100 URLs), eliminating subjectivity and becoming the "single source of truth" for quality assurance before production.