Atria
A clearer picture of everyday health.


The challenge
Workout and nutrition data live in different places and must remain useful offline. Personal recommendations need to be grounded in the user's own records, with careful boundaries around medical advice.
How I approached it
Built a Flutter client with Riverpod and Drift, a NestJS API with PostgreSQL, and shared nutrition and training calculations. Added offline workout logging, a multilingual food diary, encrypted progress photos, health-platform sync, and an assistant pipeline currently using a mock model provider.
What came out of it
A working seeded product spanning workouts, nutrition, progress and an assistant. AI responses currently use a mock provider; store purchases require production store configuration.
How the pieces connect
Offline workout runner and sync
Multilingual food diary, recipes and shopping list
Grounded assistant pipeline with a mock model provider

Next destination