We designed and implemented an AI-powered supply chain system that connected forecasting, inventory, and supplier operations into one streamlined ecosystem
• Built LSTM-based forecasting models in TensorFlow to predict ingredient demand using historical sales, weather data, and promotional events
• Integrated three separate ERP systems via a custom Node.js middleware layer, enabling unified operations across all franchises
• Used Apache Kafka for real-time stock updates and event-driven processing, reducing restock delays
• Developed an automated ordering engine in Python (Django) that triggered on-demand restocking with FMCG suppliers — as fast as 3-hour delivery
• Forecasts updated every 15 minutes, enabling dynamic response to changing demand
• Standardized data formats and ensured secure franchise-wide data exchange, fully compliant with GDPR and PCI-DSS