Working on
Stage 4a of an active rebuild. Six-product property-tech platform for a private client, solo. Frontend-implement kickoff on the Passport public-share screen this week, then remediation across 14 Stage 4a domains before Stage 3 starts. The April 20, 2026 QA run produced 53 open issues. Working through them.
Building
Methodology is locked Source-First as of April 10, 2026. Tests are regression locks, written after the QA auditor finds bugs worth locking. That shift happened after 10 weeks of Domain-Rules TDD produced a lot of clean specs and nothing shipped. A different project of mine had already shipped a zero-bug production SaaS with a simpler quality system, and that was the deciding data.
Stuck on
How to make /frontend-implement reliable enough that one agent can take a wireframe and a set of tRPC endpoints and produce a production-grade React page without a seven-round human review loop. I have no answer yet.
Available for work
Code-Rescue has one active client and room for one more. Best fit is an AI-augmented rebuild, architecture review, or long-running methodology work. max@code-rescue.com.
Where
Tampa, Florida. Working remotely.