CASE FILE // MOBILE / AI / COMPUTER VISION
FAMILIAR.
IN DEVELOPMENT
ROLE — SOLO DEVELOPER & PRODUCT DESIGNERTIMELINE — 2026 — PRESENTSTATUS — IN DEVELOPMENT
Gamified study companion (working title) — point the front camera at yourself during a focus session and on-device computer vision watches for signs of distraction: looking away, leaving the desk. Stay focused and your Familiar — a creature hatched from an ember seed — grows alongside your study streaks, earning wax-seal achievements. Offline-first, with a Firebase sync plane.
HIGHLIGHTS
- ▸ On-device distraction detection: front camera stream + Google ML Kit face detection
- ▸ Virtual companion creature rendered with Flame — grows through accumulated focus time
- ▸ "Ember seed" hatch mechanic and collectible wax-seal achievements
- ▸ Offline-first local persistence on Isar with Firebase Auth / Firestore sync
- ▸ Design-token system as single source of truth — aged-paper palette, Fraunces + Inter
ARCHITECTURE
- ▸ Feature-first Flutter architecture: core (theme / routing / constants) → features (home, session, familiar, hatch, rewards, cv) → services (firebase, persistence)
- ▸ CV service isolates the camera + ML Kit pipeline from session scoring logic
- ▸ Riverpod state graph bridges session telemetry into creature growth and rewards
- ▸ Web3 collectible layer planned behind its own interface — swappable, never load-bearing
STACK
MOBILE
- Flutter
- Dart
- Riverpod
- Flame 2D
CV
- Camera stream
- Google ML Kit Face Detection
DATA
- Isar (offline-first)
- Firebase Auth
- Cloud Firestore
- Cloud Functions