Projects

CoFiscal (Loan Default Analyzer)

  • Developed a Web App to assist borrowers make optimized loan decisions based on their predicted default probability.

  • Built LightGBM model to predict the probability with 92% accuracy after training on 250,000 data points pre-SMOTE.

  • Used Google’s PaLM-2 to provide insights into personal default cases and PDF-miner to reduce user input by 75%.

  • Led team of 4 in developing a project in 36 hours, Won Capital One Best Financial Hack Award at HackGT X.

Devpost Link

College Scheduler App (for the CS 2340 course at Georgia Tech)

  • Developed a student scheduling mobile application using Android Studio, featuring functionality for tracking assignments, tasks, exams, and classes.

  • Integrated a push notification system to provide daily alerts and reminders about upcoming due dates.

  • Utilized SQLite for database management, enabling comprehensive Create, Read, Update, and Delete (C.R.U.D) operations.

  • Implemented advanced features including prioritization based on due dates and customized weighting of exams and assignments, tailored to individual student performance.

GitHub Link