We operate 280 retail stores across Central Europe. Our forecasting is done in Excel by regional managers. It's wrong. Not slightly wrong — expiry write-offs cost us €2M last year, and out-of-stock events cost us another €3M in missed revenue.
We need an ML engineer to build a proper demand forecasting system — product-level, store-level, accounting for seasonality, promotions, local events and weather.
Stack: Snowflake for data, dbt for transformations, we're open to LightGBM, Prophet, or neural approaches — whichever fits best. You'd own the full pipeline: data prep, model training, evaluation, and integration with our replenishment system.
Requirements
– Proven experience with hierarchical or multi-store demand forecasting
– Solid Python and SQL — our datasets are large
– Experience with seasonal products and promotional lift modelling
– Has integrated forecasting output into operational systems (not just notebooks)
– Can communicate results clearly to non-technical buyers and merchandising teams