Training AI with Artificial Data: Advanced Concepts and Applications with Synthetic Data

Synthetic data—info created by programs or AI models—has become core infrastructure for how AI systems are built and maintained, but generating it without a clear purpose leads to datasets that look right and perform wrong. In this course, Tom Themeles walks you through a real scenario, from scoping your requirements to generating, refining, and evaluating structured data, and feeding it into an AI system. By the end, you'll be equipped to scope, generate, and govern synthetic data for any AI system your organization builds.


Learning objectives
Scope synthetic data requirements using the domain-purpose-method framework.
Generate structured synthetic data through effective prompt engineering.
Refine and expand datasets through iterative, conversational generation.
Evaluate synthetic data against fidelity, utility, and privacy criteria.
Train and evaluate a classification model using synthetic data.
Identify governance, security, and regulatory considerations for synthetic data pipelines.