Developed and delivered AI Data Type Subtypes Support for the ecmwf/metkit repository, extending obstype.yaml to accommodate a range of new AI-related subtypes. This work enabled robust ingestion and representation of diverse AI data types, aligning the data model with evolving analytics requirements. The implementation focused on disciplined configuration management using YAML, ensuring traceability and version control through clear commit practices. By integrating AI data-type handling into the existing pipeline, the changes prepared the system for future AI analytics workloads. The approach demonstrated careful data modeling and a methodical update process, with all enhancements delivered in a single, well-documented commit.