Congratulations! You have deployed your AI models to the Edge. How do you make sure they are performing the way they were intended to? If they are not, what can you do about it?
In this session we will deep dive into capturing observability data on Edge deployments even when the network connection is intermittent or has limited bandwidth and have the ability to return the data during specific time periods to run observability on this data.
We will show how Data scientists in the Model Operations Center can monitor drift for models deployed for a specific edge location or a group of edge locations and take action on underperforming models by hot swapping in better performing models.
Try this Computer Vision model and other common AI use cases using the Wallaroo.AI Azure Inference Server Freemium Offer on Azure Marketplace (aka.ms/Wallaroo-Inference) and also try the Free Wallaroo.AI Community Edition (aka.ms/Wallaroo.AI-Free)
#microsoftreactor #wallarooai #ai #edge #aimodel #datascience
[eventID:22425]