Resumi
Personal Project (June 2024 - Dec 2024)
Mobile app for students to build portfolios of extracurricular activities and track academic performance. Features k-means clustering to analyze activity similarities and provide personalized insights.
Architecture
Frontend: React Native for cross-platform mobile (iOS/Android), offline-first with local caching
Backend: GraphQL API via AWS AppSync with real-time sync, AWS Lambda for business logic
Auth: Amazon Cognito OAuth 2.0 for authentication and session management
Database: DynamoDB for scalable storage with optimistic UI updates
ML: k-means clustering algorithm (scikit-learn via Lambda) analyzing activity vectors based on type, time commitment, leadership level, and impact
Technical Approach
Implemented AppSync subscriptions for real-time data sync across devices with optimistic UI updates for perceived performance. Built offline-first architecture with queue-based sync when connection restored.
Moved ML clustering to serverless Lambda functions to avoid resource-intensive operations on mobile devices, with result caching for fast retrieval.