FIS Global
Customer Case Study
FIS Global provides back-end financial services for banks, credit unions, and payment processors such as Visa and Mastercard. A Fortune 500 company, FIS Global serves more than 20,000 clients in more than 100 countries. Its technology helps move more than $9 trillion around the world annually.
Industry
Vertical Use Case
- Digital banking services
- Gather insights across conversational customer channels
Technical Use Case
- Data ingest, ETL
- Machine learning
Customer TestimonialThe Challenges
FIS Global is one of the largest fintech providers of backend products for banks, credit unions, and payment processors. One of their products, Digital One, enables financial institutions to create an omnichannel experience across digital channels such as web, mobile and conversational channels. The platform also provides financial institutions with powerful customer analytics. With data volumes reaching petabytes in size, the team at FIS Global ran into issues scaling their supporting data and ML architecture:
- Massive Volumes of Data: Customer banks generate up to nine times more data through online and mobile than in the physical branch. This resulted in hockey stick growth of semi-structured data making it extremely difficult and cost prohibitive to scale legacy data stores.
- Complex Data Lake Infrastructure: A data lake stored in legacy infrastructure on expensive hardware was time-consuming and costly to manage.
- Inefficient Machine Learning: The FIS Global team was investing in machine learning to enhance the platform’s capabilities, but experienced challenges standardizing the storage, management and development of models across data science teams.
The Solution
FIS Global leverages the Databricks Unified Analytics Platform to reduce the DevOps overhead for managing infrastructure and streamline the machine learning process to make it more trackable and reproducible.
- Automated Infrastructure: Fully managed infrastructure powered by Apache Spark with automated cluster management for speed, cost control and elasticity.
- Delta Lake: FIS Global’s data lakes now have reliability, security, and performance, enabling engineers to build robust streaming data pipelines at scale.
- Managed MLflow: With MLflow, FIS Global can easily manage the entire machine learning lifecycle, from tracking experiments to monitoring production models.
The Results
With Databricks, FIS Global has built a petabyte-scale data and ML platform that supports the growth of their Digital One product while reducing operational expenses and complexity.
- Increased Productivity: Data science teams can collaborate and experiment more rapidly with machine learning. In fact, the team built a ML model to predict user churn in just two days.
- Lower Operational Costs: With automated infrastructure and streamlined processes, FIS Global has reduced operational expenses by millions of dollars while improving project success.
- New Products: With less time spent on infrastructure and model management, teams are now more focused on analyzing data and as result, have accelerated new product development.
“With Databricks not only are we able to tackle million-dollar projects at a fraction of the cost, we are able to be highly successful.”