Two live products. One operator. Open the front door.
Nexural live
Sage Ideas Academy live
This site is a work sample.
The thing I sell is "ship it, then prove it works." So this site does exactly that to itself: every deploy runs a Lighthouse quality gate and an accessibility scan, it's fully localized in three languages, and the client portal, proposals, and payments are real wired systems — not mockups.
Runs its own QA gate
Unit tests, Playwright smoke + a11y (desktop & mobile), and Lighthouse SEO / a11y / best-practices thresholds — all enforced in CI on every build. The same red-means-red discipline I bring to your repo.
See the live CI runs →Interactive demos, live
An AI grader and an AI receptionist you can try in the browser — not screenshots.
Try the demos →Real eval engine
Run an LLM-as-judge evaluation and read the scored output yourself.
Open the eval demo →Code you can read, not just claims.
playwright-sdet-regression-suite
A Playwright regression suite that runs 37/37 green in CI in ~15s against a public demo site — the pattern I bring to flaky test estates.
View on GitHub →llm-eval-gate
A starter eval gate for LLM output — the skeleton of the citation/quality gates I build into RAG systems.
Browse my GitHub →50+ original repos
The full body of work — platforms, tooling, and experiments — is public. Look around.
github.com/JasonTeixeira →Measured outcomes, with the receipt.
The QA platform that refused a fake green
A CVE gate went red mid-release, blocked the ship, and was back to 13/13 green within the hour.
A suite a release manager can trust
Playwright regression specs green in CI, in seconds, with zero flakes.
An assistant that can't answer without a citation
A RAG dashboard that enforces citation coverage on every response.
Full write-ups, including prior full-time roles (HighStrike, The Home Depot) with self-reported figures clearly labeled, are on the case studies page.
Bring the AI feature you're not sure about.
A free 15-minute call. You leave with a plan whether or not we work together.