AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Process is Key in Data Productionization
Process is fundamental in determining if a piece of code or data becomes part of a production system. Characteristics that define data as production-grade are trustworthiness, clear ownership, understanding the data meaning, mechanisms for data evolution, and the ability to manage context iteratively. Moving from experimental to production data requires following processes such as Continuous Integration/Continuous Deployment (CICD), unit tests, integration, and contracts. Data must have a contract and go through a productionization process to be considered production-grade, warranting different environments for such data assets. The absence of a data contract and completion of the productionization process restricts the use of the data in activities like creating machine learning models or sharing with the executive team for reporting purposes.