Short Term Load Forecasting with LSTM/CNN

Led the development and deployment of a short-term electric load forecasting system using deep learning models (LSTM and CNN) for the national grid. The solution was implemented in Databricks and Azure Machine Learning, integrating scalable MLOps pipelines for automated model training, validation, and deployment. The project improved forecast accuracy and operational reliability, enabling data-driven decision-making for energy management. Responsibilities included end-to-end pipeline design, data preprocessing, model optimization, and collaboration with multidisciplinary teams.