Data Scientist, Demand Forecasting

Pantry Loop helps 450 food banks across Europe track donations, stock and distribution. The hard problem is timing: fresh food arrives unpredictably, and need spikes at the end of each month and during school holidays.

You'll build the forecasts food banks use to plan pickups and volunteer shifts. In the first 90 days you'll audit the current model (a seasonal baseline that's off by 30% in December), ship a hierarchical forecast per food bank and category, and design a way to explain each forecast in plain language to warehouse coordinators. You'll use Python, statsmodels, LightGBM and dbt.

We're a 32-hour-week company: Friday is off for everyone, at full pay. The team works within European time zones, writes things down in shared docs, and meets twice a week.