Back to all projects
Defense / Enterprise

AI-Based Inventory & Logistics Optimization

Predictive analytics for inventory forecasting and logistics optimization in large-scale operational settings with complex multi-echelon demand patterns.

Role Developer
Domain Supply Chain / Forecasting
Status Open Source

Overview

An inventory and logistics platform that forecasts demand, recommends reorder points, and optimizes distribution across a multi-echelon network. The system combines a modern web application for day-to-day operations with a predictive analytics layer that learns from historical movement data to reduce stockouts and carrying cost simultaneously.

The Problem

Large operations carry thousands of SKUs across many locations. Simple reorder rules either over-stock safe items or under-stock volatile ones. Demand correlates across locations in non-obvious ways, and lead times vary. Teams end up firefighting rather than planning. A learned forecasting layer, coupled with clear operational tooling, turns this into a tractable problem.

My Role & Contribution

  • Built the full-stack web application for inventory, orders, and transfers
  • Implemented the demand-forecasting layer and integrated it with the operational workflow
  • Designed the dashboards and reporting used by operations staff

Approach

  • Node.js / Express backend with MongoDB for inventory, orders, and transaction history
  • React front-end providing role-based views for operators, managers, and auditors
  • Python analytics service for demand forecasting (scikit-learn + time-series models) exposed via a simple REST interface
  • Reorder-point and safety-stock computation driven by forecast uncertainty, not fixed rules
  • Reporting and alerting for slow-movers, stockouts, and anomalous consumption

Tech Stack

Python JavaScript Node.js Pandas scikit-learn Time-series forecasting MongoDB React.js

Results & Impact

  • Forecast-driven reorder recommendations replacing fixed-rule logic
  • Open-source codebase with documentation for others to fork and extend
// TODO: add diagrams / screenshots
← Previous Predictive Maintenance — WLR Swathi