Astreia is a machine learning model verification system for public and private use.

Its primary purpose is to serve Model Users who need a machine learning (ML) Model to generate predictions to fulfill use cases such as forecasting, decision support, or product recommendation. The fundamental assumption made my Astreia is that Model Users do not have the technical capacity to build these systems, so they need to contract Model Providers, that is, third party software services.

Astreia aims to solve three fundamental problems in real-world applications of machine learning:

1. Model Users are often unable to assess the quality or health of Models offered by Model Providers.
2. Models degrade in quality over time due to covariate shift.
3. Model Providers are insufficiently incentivized to maintain the quality of their Models.