Synthetic Data Generation with AI

Creating realistic test data for development environments has always been a pain. You need enough volume to be meaningful, enough variety to catch edge cases, and it needs to actually look like your production data - without being your production data. AI can help with this, and we've got options for how to approach it.
This session explores using AI to generate synthetic data for your data platform. We'll look at different approaches - from simple setups using VS Code connected to Fabric notebooks, to building dedicated MCP servers that give us the options for agentic creation. Each approach has trade-offs, and we'll dig into when each makes sense.
We'll see this working in a live demo where we generate datasets on the fly and show how quickly you can build simple test environments.
Expect to learn:
Different approaches to AI-powered synthetic data generation
Using VS Code with Fabric notebooks for quick data generation
Building MCP servers for more controlled, repeatable workflows
Real examples of synthetic data for development environments
How to make generated data realistic enough to be useful
When each approach makes sense for your situation
This session is for anyone managing development environments who's tired of fighting with test data. It's practical - expect to see working examples and leave with approaches you can implement.

Creating realistic test data is a pain. AI can help. This session shows different approaches - from VS Code with Fabric notebooks to MCP servers - for generating synthetic data. Live demos, practical examples, real solutions.