Tej Pandey

Enhanced and deployed an end-to-end machine learning application for retail demand forecasting using the M5 Forecasting dataset. Built FastAPI REST APIs for health monitoring, single prediction, and batch prediction; expanded a React and Vite frontend with JSON/CSV batch forecasting and session-based prediction history; deployed the frontend on Vercel and backend on Render; and implemented Vitest frontend testing with 20 passing test cases covering prediction workflows, validation, and API integration.

Developed a machine learning model using Natural Language Processing to classify news articles as real or fake. Applied data preprocessing, feature extraction, and model training techniques, implemented Logistic Regression and Naive Bayes classifiers, and evaluated performance using standard classification metrics.