Fast and Cost Effective Machine Learning Deployment with S3, Qubole, and Spark
In this session, join Piero Cinquegrana, Senior Scientist with Marketshare, a Neustar Solution, for a deep dive on Marketshare’s Machine Learning projects. You’ll learn:
- Why all companies that work with Big Data need fast and cheap storage such as S3;
- How AWS, together with Qubole, enables decoupling of storage from CPU and memory, giving flexibility in trying different cluster configurations for Big Data analysis;
- How Marketshare combines SparkSQL connectors with Qubole to handle a variety of data formats used in S3: text, json, hive tables and others
You’ll also learn how to set up your own Spark Notebook using Qubole Data Service to build interactive and collaborative dashboards. Alex Aidun, Training Engineer at Qubole, will also demonstrate how you can leverage MLlib algorithms such as classification, regression, clustering, and more.
Lastly, Brandon Chavis, Solutions Architect with AWS, will review recent Amazon S3 feature enhancements and upgrades, including storage classes, cross-region replication and an improved consistency model.