Regular analysis of processing of precursor data through AP pipelines, including identifying issues, tracking them down, and resolving them. Includes testing algorithmic changes on large datasets as well as developing and testing time series features.