GET /work/focusflow-ai
FocusFlow AI
Productivity platform with a React Native app, a Node/Express API, and a separate Python service that runs 3 scikit-learn regression models to predict study schedules.
ARCHITECTURE
A React Native (Expo) app talks to a Node/Express API secured with JWT. The API stores data in SQLite and works with a separate Python service that runs three scikit-learn regression models to predict study schedules.
SECTION /decisions
Key Decisions
Three runtimes, one product
The system is deliberately three parts, not one: a React Native (Expo) app for students, a Node/Express API that owns auth and data, and a separate Python service that runs the machine learning. The Express API stays a focused CRUD-and-auth service, and the prediction logic lives behind its own service boundary where Python and scikit-learn are the right tools.
Predictions as a service, JWT and SQLite for the rest
Predictions come from three scikit-learn regression models that predict study schedules. Auth between the app and the API uses JWT, and SQLite is the datastore.
SECTION /challenges
Challenges
The hard part of a three-runtime system is the seams: keeping the mobile app, the API, and the ML service in agreement about data shapes while each side evolves.
SECTION /result
Result
A working end-to-end platform: a React Native app backed by a Node/Express API, with a separate Python service generating predicted study schedules from three regression models.