No Chip on the AI Shoulder
AI models have no guilt, no wound, and no chip on the shoulder to steer them. They work without the psychology we use to explain each other, which is why their behavior still feels so foreign.
Read More →
My engineering background brings practicality, my artistic background brings whimsey, and my QA experience brings clarity. Those qualities converge on a sense of things that people actually want to use, like an environmental reporting app, a school for Quality Engineers, or a casual drawing game you can share with a friend.
Quality engineer by trade. Builder, musician, and dreamer the rest of the time.
Things I have concieved, built, shipped, and maintain.

Scan a product and get an AI-generated environmental impact score, plus earth-friendlier alternatives. Now on the App Store.

A school for testers: practical, human-verified lessons for people entering QA and for engineers sharpening the craft.

A daily drawing game with generated starters and tools. Sketch on your own, or draw together with a friend.

A Chrome Extension for LinkedIn that helps job hunters organize and filter their search results.

Bridges Cypress and Playwright test runs with Google Sheets or CSV, streamlining test result integration and boosting team collaboration.

A social network tailored for musicians and music professionals. Showcase industry skills, find others and connect through events and groups.

A dog walking and pet sitting service that provides a user-friendly platform for estimating and booking pet care services.

A community-powered charity that aims to enrich the lives of learners through the funding of musical instrument supplies.

An interactive simulator that models how societal factors like currency inflation, tax rate, and housing cost impact social classes and overall stability.

A website for a civilian training, survival, and veteran rehabilitation facility

An online sales portal for renting audio equipment and labor for live events and studio sessions.
I spend my days asking quality-questions about test automation, QA leadership, product development, and making sure the things we ship still work for the people who use them. Off the clock I build products I actually want to live with: Blue Dot on the App Store, QA Academy for testers finding their footing, Doodle Do for drawing with a friend. A background in music still sneaks into how I work: listen first to understand the audience, then make them feel something that holds an experience together.
Remote · Feb 2026 – Aug 2026
Owned production triage and deploys—built AI agents to automate issue routing and release workflows while protecting pipeline stability and uptime.
Remote · 2022 – 2025
Led JavaScript automation at scale — created peer review standards and maintained resilient E2E suites.
Remote · 2023 – present
Founded qa-shadow-report — translating Cypress & Playwright results into spreadsheet reports for team-wide test visibility.
Hybrid · 2021 – present
Appraised methodologies, risk, and technologies to guide a company through Series A funding.
Hybrid · 2019 – 2022
Delivered client features while expanding test coverage and contributing to engineering and design.
See how my skills and experience can make a positive impact on your project.
Leverage cutting-edge AI to automate testing, optimize user flows, and drive efficiency, saving time and reducing human error.
Ensure robust, accessible, performant apps with thorough validation that catches issues early.
Build modern, scalable, SEO-optimized web apps from scratch or integration—with reliable long-term support.
Lead teams to ship faster through joint planning, resource allocation, and collaborative processes.
End-to-end ownership from ideation to launch with startup efficiency and impact focus.
AI models have no guilt, no wound, and no chip on the shoulder to steer them. They work without the psychology we use to explain each other, which is why their behavior still feels so foreign.
Read More →Two different failures get treated as the same problem: the model never had your idea, or it had the idea and then wandered. The work is briefing like a PM, checking like a tester, and asking the model to examine itself before you trust the result. A failure to understand the model is a failure to understand the work.
Read More →In interviews, skip the fancy stack and buzzwords. A strong grasp of the fundamentals makes a stronger impression—and it is what actually transfers to any tech stack.
Read More →Please contact me directly at petermsouzajr@gmail.com or through this form.