Client's challenge
This client is helping tens of thousand of consumers all across the globe with heating, ventilation and air conditioning (HVAC) appliances. These appliances collect sensor level and metadata each second.
The client has currently two (patented) machine learning use cases. One involves an energy savings calculation model and another is for detecting leaks/monitoring the general health of a specific set of appliances.
The challenge for the specific customer was to enable and deploy the right machine learning models in their production environment
The Solution
We set up and deployed an Azure Machine Learning (AML) platform in combination with a third party feature store and several Azure services, to meet monitoring capabilities. All the required environments, compute resources, pipelines, model registries and endpoints were prepared and deployed to AML. The necessary guidelines to ensure a reliable, efficient and good model building practices have been followed.
Created impact
Thanks to our Azure Machine Learning platform, the two patented Machine learning models are now available in a production environment and allow for time reduction in future model development.
The client now has all the tools and processes in place to perform MLOps, and efficiently serve/track models to server business and consumer needs in an agile way.
