Can you predict the stock and bond prices? This is one of the most difficult topics in Artificial Intelligence and Data Analysis. Niels Denissen of ING will tell about there cutting edge technologies they use to help traders make their decisions, and the problems they are still facing.
Katana is a streaming platform designed for traders working at ING. It aggregates and enriches multiple financial market sources in real time to provide a single overview for them to base their decisions on.
Additionally we use machine learning to augment this process by providing predictions with the knowledge of hundreds of thousands previously traded deals. With Katana, traders are shown to be faster and more accurate in their decision making process. Katana has been built with open-source software, including technologies such as Kafka, Flink, CouchDB, and Docker. The analytics stack has been built using Python, using Pandas, SK-Learn, and Keras.
My name is Niels Denissen and I currently work as a Data Engineer in ING's Wholesale Banking Advanced Analytics team for project Katana. Before starting at ING roughly 2,5 years ago I finished a masters in Artificial Intelligence at the University of Utrecht and a bachelors of Computer Science at the Technical University of Eindhoven. I'm passionate about working with data: Providing fast and reliable solutions for complex problems, as well as practicing data science whenever time allows. Outside of work I like cycling, soccer, snowboarding, and value time spent with friends and family.