Data Observability for Machine Learning

Rapid Verification

Machine learning model verification is essentially a data pipeline. Once a model is trained, a common question is, “How does this version of the model perform vs a previous version of the same model?”

Validator can help ensure the consistency of both:

• the input to the model
• the model output as it corresponds to the model’s inputs

You can send Validator a variety of data types for inputs, and send data from multiple stages, ensuring that “preprocessing” into your model also behaves correctly. This rich, multi-stage data capture enables not only rapid model verification but also verification of the supporting code around the model’s deployment.