CoachFit / FitIn
A mobile-first AI coaching product I architected and shipped solo for a small founding team — from empty repo to a polished MVP with real-time features, AI personalization and monetization.
Solo MVP, end to end
I architected and shipped the complete application: a React Native (Expo) app backed by a RAG pipeline combining the OpenAI API with Supabase vector search, which cut hallucinations significantly and kept coaching answers grounded.
I owned everything — data model, auth, real-time chat, UI system and release — collaborating directly with the founders on scope and priorities.
Real-time, low-latency backend
Supabase Edge Functions handle authentication, subscription management and real-time chat, keeping critical paths fast. I added WebRTC and TTS for live video coaching and voice-guided workouts — clear differentiators in a crowded fitness market.
Monetization & design system
I implemented a freemium model with Stripe payments, native In-App Purchases (iOS/Android) and AdMob to maximize revenue across tiers, and built an atomic design system with dark/light mode that cut UI feature dev time substantially.