Raiffeisenbank: A Unified Data Platform Concept on Databricks in AWS Cloud
From migrating core technologies to the systematic development of analytics, AI, and cost management
use cases delivered in the first year of platform operation
users involved in the first community meetup
APEX platform as the standard for analytics development across the bank
Challenge
A Unified Data Platform for Analytics, AI and Cost Management
Raiffeisenbank aimed to build a unified cloud-based data platform for the entire banking ecosystem and to establish APEX as the standard for analytics development. To establish a solid technological foundation for the further development of data, AI, and MLOps, the bank turned to Adastra, which brought hands-on Databricks expertise to the project and enabled its adoption within the bank, where it is now used by internal teams.
In parallel, the bank, in cooperation with Adastra, required the implementation of transparent cost allocation across teams, the recommendation of a suitable ETL tool, and the preparation of an outsourcing study, including an exit strategy.
Solution
Migration to Databricks on AWS and Deployment of the First Production Use Case
The foundation of the solution was the migration from legacy technologies such as Hadoop, SAS, and existing data warehouses to Databricks on AWS, complemented by a robust governance layer. We subsequently refactored data pipelines and ML workflows with a strong focus on scalability and the long-term sustainability of use cases. Based on an in-depth understanding of the bank’s environment, we recommended the ETL tool best suited to its specific needs.
By implementing the first production-grade use case, we validated the entire architecture in practice and ensured the approval of security controls. Throughout the preparation and rollout of the reference use case, we systematically transferred Databricks know-how to the bank’s internal teams.
Impact
A Data Platform Ready for Real-world Banking Operations
The outcome of the project is a modern data platform built on Databricks on AWS, which Adastra fully adopted into the bank’s environment, onboarded the first teams to, and ensured adherence to current best practices. Today, the bank operates a secure environment usable from PoC through to full-scale production, with GenAI and MLOps tooling deployed in compliance with regulatory requirements and cybersecurity standards.
The project established a strong foundation for the systematic development of analytical use cases across the bank, as well as for cost management and transparent chargeback. The deliverables also included an outsourcing study with an exit strategy, providing the bank with long-term technological and strategic confidence.
