“Serving AI with Data” – Recap of the Microsoft AI Immersion Workshop

Artificial Intelligence is enjoying its day in the sun. AI-related articles are now routinely featured as front page news and the technology is being used in an ever growing variety of applications, including in conversational bots, autonomous vehicles, connected machines, in medical diagnosis and much more. AI is therefore of growing interest to millions of developers worldwide, and to companies of all stripes, from garage startups to the largest enterprises. Microsoft has a long-term commitment to AI, and, working closely with developers, partners and customers, we wish to democratize AI and bring its benefits to every person and every organization. Not surprisingly, AI was a central theme at Microsoft Build earlier this month and, based on the feedback we received from attendees and in media reports such as this one, our vision and message seems to be resonating with the broader community.

For the vast majority of developers, however, it's still early days when it comes to skilling themselves on AI, and certainly the bewildering array of AI tools and technologies out there can make things seem a bit daunting. When we decided to host our first-of-its-kind Microsoft AI Immersion Workshop earlier this month, our goal was to give as many developers as possible an opportunity to get their hands dirty, write code and build apps using Microsoft's AI and intelligent data platform stack (which includes may popular open source packages too, btw!). The Workshop was offered as a free pre-event to Build 2017 and was later opened to all developers outside Build as well.

This blog post is a short summary of the Workshop, and in it, we've included key resources from the Workshop.

Kicking things off, Joseph Sirosh, Corporate VP for the Microsoft Data Group, delivered a keynote talk, Serving AI with Data, to a packed house at the W Seattle. In his talk, he describes how the confluence of the cloud, data, and intelligence are driving a new AI revolution, one that could rival the industrial revolution in terms of its transformative potential. Joseph's keynote included many cool demos and customer success stories, and was followed by a very popular open Q&A session with him. Catch his talk, demos and Q&A via the video link below – it's a great way to get an overview of Microsoft's comprehensive approach to this space, and also how it will benefit you, the AI developer:

Following Joseph's keynote, we ran five parallel workshops, on the topics listed below. Each of these workshops lasted six hours (no less!) and featured hands-on technical learning, led by Microsoft domain experts. Materials pertaining to each workshops are posted to this GitHub site and we hope you get a chance to browse through the same.

Applied Machine Learning for Developers Workshop focused on Microsoft R Server and Azure Machine Learning. Brief introduction to the R language and IDEs for R, followed by R in the cloud and deployment option use cases. Basics of the data science process, from identifying the question that you would like to see answered, to determining which algorithm to apply, to evaluating how good your model is. How to use Azure ML to apply the data science process and build an ML model, and deploy it as an API on Azure.
Big AI – Applying Artificial Intelligence at Scale Workshop focused on applying AI at scale using a variety of technologies. Topics covered include planet-scale AI on structured and unstructured data, deep learning at scale using Azure, and intelligence at scale using Azure Data Lake.
Building Intelligent SaaS Applications Workshop covering the end-to-end flow of building intelligent Software as a Service (SaaS) apps in Azure. Combining presentations and hands-on experiences, this covers the data tier, including different SaaS design patterns to consider when building, managing and monitoring a scalable and cost-efficient multi-tenant data tier; the app tier, featuring the powerful AI technologies available for you to build intelligent apps; followed by the analytics tier, showcasing how to power interactive analysis on your app telemetry data, backing your SaaS app with a highly scalable data warehousing solution.
Deep Learning and the Microsoft Cognitive Toolkit Workshop focused on deep learning. Introduces deep learning and common use cases, including a hands-on tutorial on how to use the Microsoft Cognitive Toolkit, followed by the different Azure services that you can use to build an end-to-end deep learning solution. The session helps attendees familiarize themselves with the different areas of deep learning, gain hands-on experience in practical applications of deep learning, and recognize the different Azure services to use, when building deep learning solutions.
Weaving Cognitive Services, Azure Services, and Bots to Provide Next-Generation Intelligence This workshop is about how to use the power of deep learning models that have been pre-trained to solve common ML problems such as face/emotion detection, sentiment analysis, and more. It starts with a guided tour of Microsoft Cognitive Services and the Azure service offerings that pair well with them. It then walks through multiple application scenarios involving integrating Cognitive Services together, as well as with Azure Search, Cosmos DB (NoSQL), and other Microsoft technologies to provide cohesive end-to-end solutions. Attendees use their learning to create a bot that uses these services to find specific vacation photos (e.g. "find photos with kids outdoors looking happy").


Attendees got tremendous value from the Workshop and had exceedingly positive feedback to share about it. Their favorite aspect of the event was the opportunity to not just see but use these technologies in real life, and with seasoned experts to guide them. The attendee quotes below sum up the day's proceedings:

"Well-planned, well-organized, very heavily attended, and provided a lot of value. Probably the most productive software engineering workshop I have attended in the past ten years or so. There was no fluff – it was all hard information and guidance."

"Presenters were very highly qualified, world-class ML professionals, but all of them were down to earth, easy communicators that spoke at my level (beginner / intermediate ML) where I could take away ideas I can use in my real day job."

"Keynote Q&A was excellent, very direct answers to a wide variety of broad questions."

"Seeing the momentum of investments and the opportunity to connect to Microsoft folks that work on the technology."

We'd like to thank all the attendees for spending the day with us, learning about AI, and sharing their thoughts. As noted above, all Workshop materials are posted to this GitHub site, and, to get a comprehensive view of Microsoft's AI and intelligent data platform approach, do tune into the keynote talk, available here on Channel 9.

We hope you will join us on our mission to democratize AI ! 

CIML Blog Team