A generic monitoring package for machine learning models developed for Rijkswaterstaat

The client,

Rijkswaterstaat has been responsible for the Netherlands’ roads and waterways since 1798. It is the executive agency of the Ministerie van Infrastructuur en Waterstaat. It manages and develops national roads, waterways and waters, and works towards a sustainable living environment. Together with other organisations, it is working towards a country protected from flooding, with enough green space and enough clean water. And where you can travel safely and smoothly from A to B.

Client Rijkswaterstaat

Industry Government

Services AI Solutions

Tools & techniques Python, pytorch, tensorflow, Alibi-detect, Pydantic

Our solution,

Rijkswaterstaat has developed and deployed several AI models. To maintain the quality of these models in production, they need to be actively monitored for issues such as data drift and prediction drift. Manual monitoring does not scale, so a generic monitoring package is needed to detect drift and outliers. We developed a generic monitoring package for all these models.

DSL came up with an inventive solution that we can put to good use.

Nanine Overgaauw Rijkswaterstaat - SRE Data Science

Cases

View other case studies too,

More examples of what data and AI make possible in practice.

A modern office building with a curved façade.

Developing an automated and efficient validation & publishing pipeline in GitLab

View case study

Discovering the ‘next Max Verstappen’ using data and smart algorithms.

View case study

What do you want data to do?

E-mail info@datasciencelab.nl Phone +31 (0)20 3636 724 Office Westerdoksdijk 599/601, 1013 BX Amsterdam