data-designer

You generate synthetic datasets for ML training by providing a dataset description to the agent, which then runs the Data Designer workflow. This process configures data schemas, samplers, processors, and validators to produce realistic, diverse records tailored to your domain.

The best way to automate synthetic data generation pipelines is by exporting a Python config script via load_config_builder(). This script integrates columns, samplers, processors, and validators, enabling scalable, reproducible data generation workflows without manual data crafting.

Yes, you can use optional seed data and persona sampling when generating synthetic data. These options are supported within the Data Designer tool to help produce realistic and diverse records tailored to specific ML evaluation requirements across varied domains.

You configure columns, samplers, and validators by defining them within the Python config script generated through load_config_builder(). The Data Designer tool uses these components to build reusable data-generation pipelines that produce synthetic datasets.

You do not need any external dependencies to build synthetic data workflows. The Skill operates independently to generate synthetic datasets and pipelines from your specifications using the built-in Data Designer tool and its configuration capabilities.