ODSC Europe 2023 Keynote: Microsoft’s Henk Boelman on Deploying PyTorch Models with Azure

As part of our ODSC Europe virtual program, we were fortunate to have Henk Boelman, Senior Cloud Advocate at Microsoft, kick off the virtual conference with the keynote talk. Henk’s specialties include AI, Azure, and application development. He is also an experienced speaker and has given addresses at user groups and international conferences.

Henk’s keynote talk, Build and Deploy Pytorch Models with Azure Machine Learning, is available for free now on our Ai+ Training Platform, which you can access with a free account. Check out our quick summary below.

The beginning of the talk lays out the current landscape of Microsoft Azure, with 8 research centers and 1,000 researchers worldwide. With 22,000 AI-related patents and more than 3,000 research papers published, Microsoft Azure has a deep well of research underpinning their technology and services.

In their portfolio, you will find services for business users and citizen developers, including Microsoft 365 and the Power Platform, as well as products aimed at developers and data scientists, including Azure Machine Learning.

Diving deeper into Azure Machine Learning, you’ll find that it is versatile and able to integrate with your existing systems. It offers the option of low-code auto ML, partial-code ML, and full-code ML. What’s more, when using the full-code service, you can use your favorite framework or library, and choose to work on the cloud or with Kubernetes.  

Additionally, Azure Machine Learning can help in the messy process of bringing a model from prototype to production. As Henk notes in his talk, although we think of the machine learning lifecycle as a linear process, in the real world, it is more iterative, requiring retraining and monitoring. 

Azure Machine Learning is designed with this more iterative reality in mind, helping you to scale your machine learning in a responsible way. 

The second half of Henk’s keynote talk consists of 3 demos, Prototype, Training, and Operationalizing. Be sure to check out Pytorch Models with Azure Machine Learning on-demand to see Azure Machine Learning in action. 

If you’re an avid fan of Azure, then be sure to check out our page entirely devoted to our partnership with Microsoft Azure here.