Simone Data Science Platform
All-in-one smart factory data science solution for Simone, Korea's leading luxury handbag manufacturer.
The Challenge
Simone's factories across Vietnam, Cambodia, and Indonesia generated enormous volumes of production and quality control data stored in disconnected systems, making it impossible to detect defect patterns or forecast production efficiency at a global scale.
The Solution
Built a unified data science platform that ingests factory data across all sites, runs ML-powered defect detection and production forecasting models, and surfaces actionable insights through an executive dashboard accessible from anywhere.
Simone Data Science Platform is an enterprise smart factory intelligence solution built for Simone Handbag — one of the world’s largest luxury handbag manufacturers, producing bags for brands including Coach, Kate Spade, and Tory Burch. The platform consolidates production metrics, quality inspection data, and supply chain signals across multiple international factory sites into a single analytical environment.
Challenge
Simone operated dozens of production lines across Southeast Asia, each with its own quality control workflows and data silos. Defects were caught late in the production cycle, leading to costly rework and waste. Factory managers lacked visibility into cross-site performance trends, and the data science team was spending more time on data wrangling than on actual modelling. A unified, production-grade platform was needed to operationalise machine learning at factory scale.
Solution
The platform is built on a Python data pipeline that ingests structured production logs, QC inspection records, and sensor feeds into a central PostgreSQL warehouse. ML models trained on historical defect data generate real-time quality risk scores for active production batches, alerting supervisors before defective units reach the next stage. A React dashboard provides role-based views for factory floor managers, regional directors, and C-level executives. Docker-based deployment ensures consistent operation across factory-site servers with varying infrastructure maturity.
Key Features
- Cross-factory data ingestion pipeline consolidating production logs from multiple international sites
- ML-powered defect prediction scoring active batches in real time to prevent downstream rework
- Production efficiency forecasting with configurable target windows and variance alerts
- Role-based executive and operational dashboards built on React with PostgreSQL-backed analytics
- Automated anomaly detection for equipment performance degradation patterns
- Interactive drill-down from global overview to individual production line metrics
- Dockerised deployment supporting on-premise factory servers and cloud-hosted environments