Created a web-based tool that uses GPT to estimate and visualize volume from textual or numeric inputs, featuring live inference and responsive design.

Developed a real-time NLP pipeline that transforms toxic social media comments into positive tone using GPT API and Hugging Face sentiment models. Won 1st place at CMU-SV Hackathon.

Optimized a bi-clustering algorithm for large-scale genomic epistasis analysis, reducing runtime from 16 hours to 10 minutes via code profiling, algorithmic refactoring, and vectorized operations.

Developed a web dashboard for classroom analytics and AI-assisted student feedback generation. Used by 200+ teachers and students across multiple classrooms.

Built a full ML pipeline for the Transporter Classification Database (TCDB), including data preprocessing, feature engineering, and model evaluation for large-scale protein classification.