How Project Health works
Project Health summarizes operational evidence across delivery, quality, site analytics, and infrastructure without replacing the underlying systems.
SignalsGitHub, Vercel, Render, Neon, project-specific quality checks, and aggregate site analytics.
Software decidesWhat data to fetch, what counts as a failure, and which evidence the investigator may access.
AI helpsOn request, it explains selected failures when deterministic signals alone are not enough.
Human decidesA person verifies the diagnosis and chooses whether code, configuration, or deployment should change.
Role & attribution
This is an applied AI product exercise, not a claim that the implementation was written by hand.
My workProduct framing, workflow design, guardrails, quality model, failure behavior, QA, and iteration.
AI-assisted workImplementation support, code generation, research support, and the bounded failure-investigation summary.