ObservabilityTestingRobustnessScaling
A research platform in production at a Philippine NGO health clinic, built end to end with the Data and ML architect who led the analysis. Guided audio sessions paired with HRV recordings (Polar H10) become precisely time-synchronized, analysis-ready data across web and native mobile.
- The approach
- A four-phase session engine timestamps every transition server-side, runs offline once audio is cached, and re-syncs on reconnect, so a dropped connection never loses a session. The validated HRV analysis (RMSSD, HR, DFA) lives in one shared layer, so web and native mobile compute the same trusted metrics.
- What it proves
- In production and self-serve: researchers, participants and clinic staff run it themselves. GlitchTip (Sentry) observability, Playwright tests behind CI quality gates, and Supabase row-level security keep it sound. Delivered end to end, solo.
- Highlights
- GlitchTip (Sentry) observability that catches swallowed failures
- Offline sessions that re-sync on reconnect
- One shared HRV analysis layer for web and native mobile
- Playwright end-to-end tests behind CI quality gates
- Stack
- Next.jsReactTypeScriptTailwindNxSupabase (PostgreSQL)PythonFastAPINeuroKit2DockerRenderPlaywrightGlitchTip / SentryCI/CD
Have a process to automate, or an AI feature to ship?
Whether it's manual work between tools that should run itself, or an AI product that needs to reach production reliably, I'll take it from audit to a robust, well-tested build, end to end or alongside your team.