BC Data Colloquium: Michael Reid

As machine learning becomes more critical to businesses’ core offerings, data scientists must adapt their tools and techniques to meet business demands. Industry data scientists can no longer run ad hoc Tensorflow or Spark jobs against local data sets; they must now handle cloud computing, high availability and disaster recovery, and real-time analysis. We present an overview of the modern data science software landscape, with an emphasis on tools data scientists can use to design a scalable and productionized machine learning system. We will also briefly look at the future of data science software, including GUI-based data science tools and stream processing platforms.

Biography: 

Michael Reid is a software engineer with a decade of experience at companies such as Hootsuite, Splunk, and Amazon. After graduating from UMBC with a B.S. in Mathematics, he started programming front-end systems for US federal government agencies before transitioning to a backend developer and moving to Vancouver. He currently is a manager for RDS Oracle, Amazon Web Services’ cloud managed Oracle database. Previously, he worked on high-throughput data pipelines with tools such as Spark, Scala, and Docker; he is also a strong advocate for continuous deployment and the DevOps movement. His current interests in data science are use of modern software tools to facilitate a “DataOps” culture, as well as looking forward to the future of streaming machine learning with tools such as Flink and Apex.