If you are working with synthetic data and would like to learn more, check out the blogpost that demonstrates how to automatically detect issues in synthetic customer reviews data generated from the http://Gretel.ai LLM synthetic data generator. Source: 11 months ago

I was chatting with the founder of GretelAI recently and learned a lot about the area of synthetic data. Source: over 1 year ago

The requirement is to build out synthetic data with very similar size and shape to our production data and ideally have a framework to do this, maybe at some level use our data and "de-productionize" it. gretel.ai looks like it may be a fit, but from what I see I need to upload production data to their environment and that's a no-go. Source: over 1 year ago

> https://gretel.ai/ Where were you the last decade of my professional life? I couldn’t find anyone to take my money for exactly this. - Source: Hacker News / almost 2 years ago

[2] https://gretel.ai/blog/how-to-safely-work-with-another-companys-data. - Source: Hacker News / almost 2 years ago

It is doable. However the main focus of MLFlow is in experiment tracking. I would suggest for you to look into another monitoring tools such evidentlyai . You can track more things than performance (e.g.data drift). Which may be helpful in a production setting. Source: almost 2 years ago

Evidently is an open-source Python library that analyzes and monitors machine learning models. It generates interactive reports based on Panda DataFrames and CSV files for troubleshooting models and checking data integrity. These reports show model health, data drift, target drift, data integrity, feature analysis, and performance by segment. - Source: dev.to / over 2 years ago