Tensorflow-accelerated Genetic Algorithms & StreamSets
Speaker:
Andrew Morgan, Aon
Abstract:
Andrew will introduce ideas around Tensorflow accelerated Genetic Algorithms, reviewing an introductory use case for classification and regression using Karoo_GP, a tool for GPU accelerated Evolutionary Algorithms in python. He’ll explain the problem, the ideas and methods investigated, the theory behind the tools, and give a working demonstration. An informal review of the comparative results against other popular algorithms is included, as well as an explanation of how easy it is to deploy the Karoo_GP trained models to Apache Spark and other SQL enabled technologies.
Bio:
Andrew is an experienced big-data scientist and platform engineer, and author of Mastering Spark for Data Science. He works in London, Dublin and Krakow as Head of Data Services for Aon, and directs a large data engineering team.
Speaker:
Cristian Varela, Aon
Abstract:
Your boss needs a new system with hundreds of high-performing data pipelines processing real-time data from all sort of heterogeneous sources. And you’ve guessed right: You’re charged with developing it from scratch, with very little budget, and a small team of data engineers. If that wasn’t enough you also have a very tight deadline. Impossible I hear? Fret not! StreamSets have Open Sourced their Data Collector which enables you to develop and continuously run streaming pipelines in minutes (forget about scheduling nightmares) as well as monitoring their throughput and performance in a single integrated interface.
In this talk we’ll introduce SteamSets’ Open Source Data Collectors and develop a pipeline to consume and analyse real-time streaming data while getting familiarized with the product features and capabilities and discuss common patterns and some tips & tricks. We’ll also introduce the concept of a MicroService pipeline and create a reusable data service.
Bio:
Cristian is a Snr. Data Architect in the Aon Centre for Innovation and Analytics in Dublin with over 20 years’ under his belt working in technical roles and data related initiatives. His Experience spans many industries including Pharmaceutical, Government, Gambling / Gaming and more recently Insurance & FinTech where he has spent the last 3 years designing Data Intensive Applications and assisting with the implementation a Hadoop based Enterprise Wide Analytics Platform. He has a personal interest in everything Python and Home Automation systems. In his own time he enjoys getting the soldering iron out and play with circuits and Arduinos when his not training for a triathlon.
