Deep Learning with Spark

Please come along to our next Meetup! We have two talks from our community on Deep Learning with Spark and as usual there will be free beer and pizza courtesy of our sponsors Capgemini so please do come along!

Machine Learning on Microsoft Azure with MMLSpark

Speaker: Richard Conway

Abstract

Much has been spoken about using Tensorflow for Deep Learning which has led to some fantastic abstractions (Tensorframes) for the Spark developer. In this talk Richard Conway will introduce a new Spark abstraction for the Microsoft Cognitive Toolkit (CNTK), a python Deep Learning library which can be run on Apache Spark. This informative talk will be demo driven and focus on some popular use cases for MMLSpark and show how you can get up and running on Azure, other clouds and local environments in minutes.

Bio

Richard Conway has over 20 years experience in writing software across a variety of business lines from finance to communication. He is currently running Elastacloud, playing a number of roles in the business, looking after logistics, finances, a little bit of sales, hiring, technical architecture and writing code.

His personal interests include architectural patterns and practices for data, and he has been interested in data science for the last five years. He uses a range of languages, currently C#, Scala, Java, Python and R (and a little Lua) but will have a crack at anything. Richard has good experience with Apache Spark and the Hadoop stack in general since the early years.

Richard has also written several books and articles and delivered several hundred courses and conference presentations. He founded the UK Azure community, IoT and Data Science Innovators and is a strong contributor to open source by cloud-enabling parts including Spark, Storm and Libcloud and a bunch of useful tools on GitHub.

He is current Microsoft Regional Director and Microsoft Azure Most Valuable Professional.

Deep Learning on Intel BigDL

Speaker: Dash Desai

Abstract

In this talk you will learn how to get started with distributed Deep Learning library BigDL for Apache Spark. You will also see demo of a Deep Learning application that uses BigDL running on a Spark cluster. The application is written to identify handwritten digits (0 to 9) using a LeNet-5 (Convolutional Neural Networks) model that is trained and validated using MNIST database.

Bio

As a technology evangelist, Dash is passionate about evaluating new ideas and trends, and helping articulate how a technology would address a given business problem. He also enjoys writing and being hands-on to create compelling demos and applications.

As a full-stack developer, he has participated in all phases of software development life cycle. He started his career working for global enterprises developing applications for the city of Nashville, TN and the state of Colorado. Since then, he has been working as an engineer and solutions architect in agile environments--for tech startups in the Bay Area in varying verticals such as VoIP, Online Gaming, Digital Health, NoSQL database, and Big Data platforms.

His passions are: high-altitude hiking, downhill skiing, travel, and photography.