Quality is a system. I build the system, and the team that runs it.

The situation

Enterprise security software is judged on consistency across a fragmented estate. The same product had to behave identically across three operating systems and the hardware variations underneath them, which multiplies the test matrix fast and makes manual coverage impossible to sustain.

What I did

I built automation suites spanning Windows, Mac, and Linux so the same verification ran everywhere rather than being repeated by hand per platform. Working this close to the OS also meant owning environment and release mechanics, not just test cases — the DevOps side of getting builds provisioned, installed, and verified reliably.

I joined as a QA Specialist and was promoted to QA Team Lead, taking on direction for the QA effort and the people doing it.

Why it matters for a leadership hire

This is the environment that taught me low-level rigour: deep platform constraints, a high blast radius for mistakes, and a matrix that only automation can hold. It is a very different discipline from web QA, and being able to move between the two is the point.

The situation

Quality needed to be established rather than improved. There was no automation foundation to extend, no pipeline gating releases, and no shared standard for what "tested" meant across web and mobile — on a product handling live customer funds.

What I did

I built the automation frameworks from the ground up and established CI/CD testing pipelines so verification became part of the release path instead of a step someone remembered to run. I then defined the QA strategy and testing standards across web and mobile so quality was consistent across surfaces rather than per-team habit.

Alongside the technical build I mentored the QA team, which is what turned a set of tools into a function that could operate without me in the loop. That organisational impact is what took me from Lead to QA Manager.

Why it matters for a leadership hire

This is the clearest evidence of the job I am targeting: take a QA function from nothing to something an engineering organisation depends on, under real financial stakes, and grow the people at the same time.

The API layer

A Python integration and endpoint framework on unittest, wired into CircleCI to run automatically on every push to a branch or to master. Fast, cheap, and positioned to catch contract and integration breaks before anything slower runs.

The browser layer

A Cypress end-to-end suite covering Chrome, Firefox, and Electron, built to run headless in CI and headed locally with IDE run configurations so a developer can debug a single failing spec without fighting the harness. Separately, a Selenium framework on geckodriver for Python-side browser automation.

The mobile layer

Mobile coverage through Appium on BrowserStack App Automate, so device breadth came from cloud infrastructure rather than a drawer full of handsets.

Why it matters for a leadership hire

A QA leader who cannot reason about the automation stack ends up approving plans they cannot evaluate. I have built these layers myself, in more than one language and CI system, which is what lets me make the trade-offs credibly and review the work of engineers doing it.

In the validation role

At Alira Health I led validation engineering for life sciences clients and built automated validation frameworks for regulated environments, introducing AI-assisted testing into a setting where every step has to be defensible. Regulated work is the hardest place to adopt new methods, which makes it the most useful place to prove they hold up.

What I build with it

Tooling that reads requirements and surfaces the gaps before code exists. Test generation driven by context instead of boilerplate. Verification steps that re-read their own output and correct it. The stack is the Claude API, Claude Vision, structured output, and agentic workflows — plus locally hosted models via Ollama with retrieval, for when data cannot leave the machine.

The proof

LANA.AI — a self-hosted assistant running entirely offline on Ollama, with persistent memory, a retrieval knowledge base over documents, and agentic task scheduling. JobFinder — a cross-platform desktop application that scores postings against a résumé and then uses a verification pass to re-read each source and correct its own scoring. Both are shipped software with installers, not notebooks.