Ten years of quality engineering, now applied to the specific failure patterns of AI-generated software. I embed in your team, find the gaps, and build the systems that keep them closed.
Get in touch ↗Every defect that reaches production sets off a chain reaction. The cost isn't the bug itself — it's the trust you spend fixing it. Below: a causal map of where a single undetected defect actually lands.
Vibe coding tools — Lovable, Bolt, v0, Cursor — generate syntactically clean, plausible-looking code at extraordinary speed. What they don't generate is judgment. They miss authentication checks. They skip edge cases. They produce architecture that holds at 10 users and collapses at 10,000. The industry has a name for this now: AI-generated technical debt at production speed.
I've spent ten years reading code for what it doesn't say. That skill doesn't expire when the code is AI-generated.
Press Space to start. Jump the obstacles you can see.
The ones you can't — those are the ones that matter.
I don't hand you a document and leave. I think through your quality problem, define the strategy, and get my hands into the work until it's running.
I've been the first QA engineer at three different companies — twice remote, at different scales — which means I've had to build quality from nothing each time. I know what a QA process looks like before and after.
At Vividly I built the first automation framework while simultaneously figuring out what AI agents could and couldn't do in a test pipeline. At TestGorilla I took 0% to 40% Playwright coverage, integrated into CI, while operating as the sole QA function. At Blackhawk Network I led a team of four and reached 84% automated coverage across critical flows.
The thread across all of it: I joined companies with no QA structure and left them with one that worked.
If you're not sure what you need — just describe what's going wrong. That's usually enough to start.