An Agentic AI-based academic monitoring and parent communication platform that combines teacher task management, student performance analysis, automated report generation, PDF creation, and parent communication through WhatsApp and Email services. The system utilizes a dual-agent architecture to ensure academic accountability while automating performance reporting and communication workflow.

Developed a Retrieval-Augmented Generation (RAG) based AI study assistant. Designed and implemented a complete RAG pipeline using LangChain and FAISS. Integrated locally hosted Ollama models for privacy-preserving AI inference.

Designed an AI-powered demand forecasting platform for retail groceries to prevent fresh produce spoilage and eliminate stockouts. Trained and tuned a GradientBoostingRegressor model with custom feature engineering (day-of-week, lag 1/2/7, rolling averages), reducing demand forecast error significantly compared to baselines. Developed an interactive web dashboard connected to a containerized FastAPI backend with automated CI/CD deployment on Render.

Developed an end-to-end Machine Learning API for GreenField Farmers Co-op to forecast next-week wholesale commodity prices (Wheat, Rice, and Potato). Implemented custom time-series feature engineering—including price momentum, rolling averages, and supply shock indices—cutting prediction error with a Gradient Boosting model.