Problem
Tutoring platforms need to understand tutor quality and predict student churn, but manual evaluation doesn't scale.
Without automated quality scoring, platforms can't identify at-risk students or provide targeted tutor training.
Approach
Built an AI analytics system that analyzes tutoring session data to predict tutor performance and student churn risk.
Used Rails for the backend API, React for the dashboard, and AWS SageMaker for machine learning model training and deployment.
Solution
An automated quality scoring system that helps tutoring platforms identify performance issues, predict student churn, and improve tutor training programs.
My Role
End-to-end development: data pipeline architecture, ML model training, API integration, and dashboard UI.
Key Decisions
Used AWS SageMaker for scalable model training
Implemented real-time scoring with Lambda functions
Designed interpretable models to help tutors understand their scores
Outcome
The system enabled proactive intervention for at-risk students and data-driven tutor training improvements.