AI-Powered Supply Chain Optimization: A Fullstack Platform Integrating Machine Learning, Real-Time Inventory Intelligence, and Proactive Delay Prediction

In today's hyper-competitive global economy, Supply Chain Management (SCM) has become a strategic differentiator

for businesses across all industries. Traditional supply chain systems suffer from manual data entry, reactive decision-making,

fragmented supplier information, and an absence of predictive intelligence — resulting in costly stockouts, delivery failures, and

operational inefficiencies. To address these critical challenges, this paper presents an AI-Powered Supply Chain Optimization

Platform that seamlessly integrates modern web technologies with advanced machine learning capabilities.

The proposed system is built on a three-tier architecture comprising React.js for the frontend presentation layer, Python Flask as

the RESTful backend application layer, and MySQL as the relational data layer. The platform incorporates Scikit-learn and

XGBoost-based machine learning models to deliver high-accuracy demand forecasting and proactive shipment delay prediction.

A smart automated reorder engine monitors inventory thresholds in real time and generates supplier-specific purchase orders

dynamically. A centralized supplier management portal enables vendor onboarding, KPI scoring, and contract lifecycle tracking.

Role-based access control secured by JWT authentication ensures data governance and operational security across all user

levels. The system demonstrates a 95% demand forecast accuracy, a 40% reduction in stockouts, and a 30% decrease in delivery

delay rates. These measurable outcomes validate the platform's capability to transform reactive, manual supply chain operations

into an intelligent, proactive, and data-driven ecosystem. Integration with Razorpay enables secure in-platform financial

transactions for purchase order processing, further streamlining end-to-end supply chain workflows.