Grace Paul

Built a stacked LSTM deep learning model in Python to forecast data center thermal load patterns, using a real-world thermal dataset sourced from Kaggle.

Designed the preprocessing and training pipeline, with model performance assessed using standard regression metrics (MAE, RMSE, R²).

Aimed at enabling predictive cooling optimization and reducing simulated energy consumption through proactive, forecast-driven cooling control in data center environments.

Developed an AI-driven caregiver workload optimization engine for elderly care using React, FastAPI, and the Groq API.

Automated intelligent task scheduling and caregiver assignment based on real-time patient needs.

Built an NLP-powered misinformation detection system with a React frontend and Node.js backend, integrating a machine learning classifier to flag fake news in real time.

Presented at a national-level hackathon at VIT Chennai, competing among 300+ teams.

Designed an AI-based waste classification web app using React and MongoDB, enabling users to identify waste categories and earn recycling incentives through an intelligent recommendation engine.

Selected for presentation at a national-level hackathon, competing against teams from across the country.