Ideally, when providing analytics as a service, the customers would all provide data in exactly the same format allowing a single data engineering pipeline to transform the data into model ready features. Unfortunately, even if customers do put the data into the same format, there will be differences in encodings and meanings of the various fields. This means that the data engineering pipeline will need significant collaboration between the data science and data engineering team to modify the pipeline. This is challenging due to the collaboration, but also the data scientists ability to perform EDA on the raw data. Even when done right, the data is often subtly different causing the data science teams to need to rebuild their modeling pipeline for each customer as well. We are here to help.