Synthetic Data Generation Pipeline overview

Designs and executes pipelines that produce synthetic training, evaluation, or test datasets — defining schema constraints and statistical distributions, configuring generative models or rule-based generators, validating that synthetic outputs match real-data characteristics without leaking PII, running bias and fairness audits, and versioning artifacts in a data registry. Produces validated synthetic datasets and a data-card. Excludes model training itself.