You are here

Google Cloud - serverless technology for data analytics and machine learning

Robert Martin
Thursday, October 12, 2017 - 12:00pm to 1:00pm
Student Center East Room 217

One of the significant challenges of processing data at scale is management and upkeep of the underlying compute and storage infrastructure.  Hadoop emerged as an innovative solution to horizontally analytic horsepower with commodity hardware.  However, at large scale, management of the Hadoop ecosystem became a substantial drag on data analytics teams.  Meanwhile, the volume of data generated by streaming events and internet-of-things devices has begun to overwhelm the ability of humans to manually analyze data.  Google's solution to manage the scale of date capture and analysis has been to go "serverless".  Solutions such as BigQuery, DataFlow, PubSub, and CloudML provide massive scale storage and compute with zero infrastructure management.  This frees up data science teams to focus on capturing data and building models.  In today's talk you will get a first-hand look Google's latest serverless technology for data analytics and machine learning.

Speaker's Bio: 

Robert Martin is a Cloud Customer Engineer for Google, working with companies from small startups to the largest enterprises to harness the power of Google Cloud Platform.  Robert has worked previously as a software developer, architect, and IT director in financial services, law enforcement, and satellite telecommunications.  Robert graduated with a BS in Computer Science from Georgia Tech and is currently pursuing an MS in Analytics from Georgia Tech.