Machine Learning
Inventory Management
Safety Stock and Reorder Point Optimisation by SKU
SKU-level reorder optimisation
Fill rate guarantee
Working capital release
The Challenge
'Just-in-case' stock immobilised liquidity, occupied warehouse space and raised storage costs. Static reorder rules failed during demand peaks or supplier delays, causing lost sales. Rule-of-thumb calculations did not reflect real market volatility.
What We Built
Analytical models dynamically calculate the optimal reorder point and safety stock level for each SKU, adjusting automatically for supplier lead time variability, historical demand patterns and seasonality. The algorithm finds the exact balance point between cost minimisation and SLA compliance.
Operational Impact
Obsolete and excessive stock converted into available working capital. Guaranteed key product availability at all times. Reduced warehouse saturation enabling more agile intralogistics flows.
Technology Used
Python
scikit-learnProphet
Azure MLPower BI
PostgreSQL
