About

FitMe is a personal project — an iOS app I built to track my own fitness and wellbeing the way I wanted: fast, privacy-first, and entirely owned by the person using it. Data stays on-device by default, analysis happens locally when possible, and no health signal gets silently shipped to a cloud AI. The interesting part isn't the app. It's what happened while building it.

I built an AI-orchestrated PM workflow — /pm-workflow — to enforce the planning discipline I kept abandoning. Then the workflow itself started evolving. Caches, eval layers, dispatch intelligence, measurement. By v7.0 it was routing to hardware-aware models.

This site is the guided tour of that process. All docs are from real shipped work. All metrics are measured, not estimated. The regressions are as public as the successes.

Links

• fitme-showcase — canonical markdown docs (GitHub)
• pm flow — the phases, skills, and shared-state model
• framework — the six-floor architecture
• design system — tokens, principles, and shipped UI
• case studies — chronological deep-dives
• research — where the framework pointed next

Contact

• Email: Regev.ba@gmail.com
• LinkedIn: linkedin.com/in/regev-barak
• Resume: Download PDF