CloudTune's goal is to invent new software systems that remove human decision-making from financial planning and scaling processes. We envision CloudTune Forecasts driving end-to-end automation that provisions Amazon's services appropriately for a great customer experience. We aim to eliminate vast quantities of undifferentiated and often tedious work for Amazon development teams - our opportunity for impact to tens of thousands of developers is enormous. The team develops large-scale scale models to inform team-level budget allocations and procurement/ allocation of compute capacity for Amazon businesses during new product launches, high velocity events and non-peak periods. This role will be contributing to a managed service that uses historical data and business signals to deliver time series forecasting for specialized use cases. The service combines a variety of distinct forecasting models, including neural networks, to produce highly accurate forecasts.