At Twinn.live, I’m contributing to Python/Flask backend development for a live-streaming platform modeled on TwitCasting. I’m also conducting D2C market research across Indian fashion brands and producing UGC video content using Google Flow/Veo.
I built a Multi-Domain Sentiment Analysis System that processed IMDb, Amazon, and news text using TF-IDF and Logistic Regression. It achieved 90.04% IMDb, 79.41% Amazon, and 88.82% news accuracy.
I built Urban Vision AI from scratch, covering dataset collection, preprocessing, and AI models for monitoring traffic, accidents, potholes, crowd density, and air quality. I created a visualization dashboard with Agentic AI and RAG for contextual retrieval, alongside automated alerts.
For my Employee Attrition Prediction Model, I analyzed an employee dataset with SQL, EDA, and feature engineering, then built a Logistic Regression model with 86% accuracy and deployed it on Streamlit.
