Kangan Sharma.

Developed a collaborative music queue system allowing users to vote and control playback order. Implemented client-server communication using WebSockets (Socket.IO) for synchronized updates. Designed room-based access control and moderation to manage multiple users and reduce spam.

Built a machine learning model using XGBoost on 20K+ records achieving 85% recall. Designed a dual-model pipeline to generate personalized health risk scores. Performed data preprocessing, feature engineering, and model evaluation. Developed an interactive interface using Streamlit for predictions and visualization.

Analyzed 3,900+ transactions using Python and SQL to extract customer behavior insights. Performed data cleaning, feature engineering, and business analysis in PostgreSQL. Built a Power BI dashboard to visualize KPIs and support data-driven decisions.