Model Versioning

The practice of systematically tracking and managing different iterations of machine learning models, including their code, data, configurations and performance metrics.

Keeping track of every version of your AI model so you can compare them, roll back if something goes wrong or understand what changed. Like version control (Git) but for AI models.

Model versioning supports governance by enabling organisations to track changes, roll back problematic updates and maintain audit trails. Your governance framework should require version control for all production AI models.