The client had time series data related to drivers and customers on a food delivery platform - I trained models that would predict the location, timing, and volume of orders made by customers, matched against forecasts for delivery driver availability, and calculated required incentives to boost driver capacity in areas and time windows where there would be a shortfall