4th #AzureGov Meetup – ML, API Management, National Day of Civic Hacking

 

The fourth #AzureGov meetup took place on May 24th at Microsoft’s K Street office.  

After a short networking session, Jack Bienko, Deputy for Entrepreneurship Education, US Small Business Administration talked about the upcoming National Day of Civic Hacking on June 4th. You can find more information about the event on Jack’s recent blog post. 

Next. Bill Meskill, Director of Enterprise Information Services, Office of the Under Secretary of Defense for Policy, kicked off a case study presentation of a Intelligent News Aggregator and Recommendation solution that his group has developed on Azure. He started out by describing the motivations for building this solution, a summary of business objectives and why it made sense to develop it on the cloud. 

 

Next Brent Wodikca and Jim Strang  from AIS presented an overview of the architecture, followed by, a brief demo.  At a high-level, the system relies on a Azure WebJobs to pull news feeds from a number of sources. The downloaded news stories are then used to train an Azure Machine Learning model -  LDA implementation of the Topic Model. The trained model is then made available as a web service to serve recommendations based on users’ preference.  There is an also an API Management facade that allows their on-premises systems ( SharePoint) to talk to the Azure hosted solution. Azure API Management allows security policies such as enforcing a JWT policy and restricting the incoming IP addresses is used to secure the connectivity between on-premises and Azure ML based solution.

 

 

Finally Andrew Weiss, Azure Solution Architect, Microsoft briefly talked about  leveraging Azure Services to help solve hackathon challenges.