Problem. Business units needed accurate demand and operations forecasts to make strategic decisions for products in EMEA markets.
Approach. Designed and deployed time-series forecasting models within AWS, applying data preprocessing, feature engineering, model selection (ARIMA, Prophet, XGBoost, LightGBM, and deep learning), and performance monitoring.
Impact. Improved forecast accuracy, efficiency, and reliability, enabling better resource allocation and cost optimization across enterprise operations.