Ryan Zotti is a data engineer at Capital One, where he focuses on putting fast, scalable, open source Big Data applications into production and in the cloud on AWS. He recently co-presented at the Fast Data DC Meet up. Ryan has experience with technologies such as Hadoop, Spark, Storm, Flink, Akka, and Kafka. In his spare time, Ryan likes to work on solving difficult machine learning problems. For example, he is currently building a self-driving remote-controlled car with a Raspberry Pi and Google’s TensorFlow, and he recently co-authored a machine learning research paper with Yale researchers that was published in The Lancet, one world’s oldest and best known medical journals.
Building Real-time Targeting Capabilities on AWS
A team of data and software engineers and data scientists at Capital One are experimenting with various technologies to enable lightning-fast promotional content that visitors will see when they visit Capital One’s website looking to apply for a credit card. In this presentation we’ll first talk about some of the technologies that we’re exploring such as the Akka-based Play framework, and H2O, a popular open source machine learning library. We will explore our evolution of data science and the H2O tools used to create the groundwork for continuous and automated testing and optimization, with the ability to scale across the entire company. Then conclude with a quick demo followed by a few tips and tricks that we learned along the way.