Ahead of the three days of the Build, Microsoft organized a few whole day sessions.
I attended one of them : the « AI Immersion Workshop » session

As the title gives in, it was focused on AI running on Azure.

It started with a keynote presenting all the available services for AI and Machine Learning on Azure.
To begin, we were presented the Cognitive Services (face recognition, image description, and so on).
You can find more on this here: https://azure.microsoft.com/en-us/services/cognitive-services/
The source code of a sample application using Cognitive Services can be found here: https://github.com/Microsoft/Cognitive-Samples-IntelligentKiosk

After that, the keynote focused on machine learning and image recognition.
A sample app on GitHub (https://github.com/Azure/Embarrassingly-Parallel-Image-Classification) shows how to use machine learning to analyze aerial images and determine what kind of terrain it is.

There was also a few customer stories about machine learning : lung cancer detection and electric pole inspection by drones.

We were reminded that SQL Server 2017 was available for a few weeks now, and that it supports AI stored procedures written in R or Python.
This is especially interesting as it is a best practice in AI/Machine Learning to put the computation logic close to the data.

After this keynote, there were a handful of workshops.
I chose to attend « Building intelligent SaaS applications »

In this workshop, we deployed a multitenant ASP.NET website that sell venue tickets, and we configured machine learning for it to answer questions like :

  • What venues are the least performing?
  • Will my future venue sellout if placed in Seattle?
  • How much a discount should I give for my venue to sellout?
  • Is there customers that go to multiple venues in a given area?

Everything we did in that workshop can be found on GitHub : https://github.com/Microsoft/WingtipSaaS