Industrial · Retail Demand & Allocation Forecasting
Role: Data Scientist (Consulting)Retail Operations · Time-Series ML · Supply ChainPrior experience — Michele Sanna at DeloitteStatus: Completed
Hierarchical SKU × store × week forecasting across 200+ stores and ~2,500 active SKUs using an N-BEATS / TFT ensemble, with attribute-based transfer learning for new-product cold starts.
- Python
- PyTorch Forecasting
- GluonTS
- Snowflake
- dbt
- Airflow
Scroll to explore →
19% → 8.4%forecast MAPE reduction on launch-quarter SKUs
−23%stock-outs on top-50 revenue SKUs
−€4.1M/yearin unplanned air-freight expediting costs
Next project
Tier-1 Pharmaceutical (via Deloitte)Next
Want this kind of work in your business?
Start with the problem, not the platform. Tell us what's slow, manual or breaking.