Flexibility – how to make data science in banking work
Banking is advancing at a rapid pace; customers are demanding a personalized experience any time any place through any device. If banks don’t keep up with this demand, many fin-tech start-ups will. To stay ahead in this race, ING has made advancing her IT capabilities in general and her analytical capabilities in particular one of her core strategies.
This strategy has led to the creation of the international advanced analytics team as a center of excellence to support all parts of the bank advance their analytical capabilities. Amongst other things, this is done by pursuing data science projects together with various business units.
In his DevOpsCon keynote Bart Buter will talk about how flexible infrastructure, software and collaboration techniques have been brought together to form the International Advanced Analytics (IAA) team. Flexibility is one of the foundational keystones to this team it motivates everything from the agile way in which the projects are conducted to the tooling and infrastructure that is used.
For instance, performing greenfield data science projects all over Europe can encounter difficulties such as project goals not fully known at start, incomplete or unavailable data and certain projected techniques might fail to deliver satisfactory results. Furthermore, the business units will have different requirements, infrastructures, native languages and cultures which need to be incorporated into the project. This requires flexibility in the way the projects are staffed, run and equipped .
To facilitate flexibility in project execution the IAA team has built a Hadoop based exploration environment for the data scientists to use. Data science is a fast moving target that requires up to date knowledge and software tools. Again, flexibility is key. However, embedding such an exploration environment within a banking IT infrastructure, strict on security and compliance, has created it’s own unique challenges and solutions that will be presented.
Last, will be a look into the future, both explaining what projects are being worked on within the bank as well as looking into how new capabilities in the Hadoop ecosystem might be used for data science.
During the talk Bart will explain the choices made while setting up the team and infrastructure and give an update on what has worked and what can still be improved.
Bart is a Hadoop specialist with a passion for data science. After finishing his master’s degree in Artificial Intelligence, he has worked within the financial industry in multiple roles; from consultant and data analyst to DevOps engineer. Besides financials, he also has a history in the software and data center industries. Currently, Bart works at an international analytics team in Germany for the ING Bank. He supports data scientists with their use cases and Hadoop based exploration environments.