In an ML/AI context, data operations (sometimes “DataOps”) refer to the practices and pipelines for handling data in the machine learning lifecycle. This includes data acquisition, ingestion, validation, cleaning, transformation, splitting into train/validation/test sets, and feeding into models. It also covers maintaining data lineage and versioning (knowing which data was used for which model), monitoring data quality in production, and dealing with data drift or shifts. DataOps borrows from DevOps principles to ensure that data management and processing is automated, robust, and reproducible. Efficient data operations are crucial for continuous training and deployment (CI/CD for ML) where new data constantly comes in and models need retraining or updating.