Requirement Data platform to be moved from hdfs to aws EC2, S3, databases, batch processing using spark (lang?) Start of data fabric, platform of data Want to take it cafeteria style model, walk and pick what data you want and use it Current is specific data shapes in S3, then to database for apps (it doesnrsquot scale, itrsquos one off) They want the future to be able to build to scale and democratic the data Spark is in clusters now and they want to move to a platform service and allow database as a service (auto load data shape) AWS is required as they want to build a platform team around this resource going forward. Spark is also required (Scala is base code, some in Python, but they want to move to Java and Scala They also want to make this self-service. Need to drive system by metadata (semantic layer, model, etc.) Abstract design for meta data systems. They are using AIRFLOW Moving toward Severless app framework.