“How to make Data Science in a Box possible”
WBAA’s Krzysztof Adamski and Rob Keevil at the biggest Kubernetes and Cloud Native technologies conference.
KubeCon+CloudNativeCon is a vendor-neutral conference gathering the attention of world-class developers, IT-professionals and C-suite technology and business leaders across the globe.
This year more than six thousands of experts gathered in Barcelona, Spain to share their knowledge, focus on innovation, and discuss on the most novel trends in container-orchestration and micro-services architectures, using open-source systems like Kubernetes, Prometheus and more.
In this context, Krzysztof Adamski and Rob Keevil gave an inspiring presentation in “How to make Data Science in a Box possible”, by showing to the audience the journey and architectural edifice behind WBAA’s Data Analytics Platform (DAP) — the project that aims to democratize Data, Analytics and Machine Learning for the whole ING.
As Kris explained, his primary motivation was to share his learnings, inspire and get inspired by the open-source community:
“We would like to open-source the DAP architecture to guide others and speed up their kick-off time. The open source community can help us improve the platform even further coming with a novel or alternative approach we might have not considered or just helping us polishing the state of the platform components”-Kris
The presentation was based on the core element that distinguishes DAP — its architectural vision, emphasizing on the compute layer, the storage and its unique security model.
Rob and Kris provided historic information about the evolution of the platform from the perspective of the current technological trends. By showing a user’s journey, they deep-dived into the platform’s architecture with focus on specific features, such as a secure desktop environment, the data discovery portal (co-developed with Lyft) and the data lineage.
“We are building the core platform features listening to users’ expectations. There is still a lot of work in front of us focused on scalable data security/governance model and productising of machine learning models. We only scratched the surface of what potentially can be done better. There is considerable time in front of us to constantly improve ourselves and challenge our own decisions on the way”-Kris
The presentation received positive feedback on how DAP could be a paradigm for organizations that need to support their analytics initiatives. The questions were related mainly to the the integration of the components that DAP was built upon.
Watch the full video below:
