Red Analytica
Back to Case Studies

Identity Verification Sector Case Study

Engineering a Scalable Lakehouse for Advanced Fraud Analytics at Jumio

How we leveraged Apache Iceberg on AWS to enhance data processing for large-scale fraud detection.

The Challenge

Jumio, a leader in identity verification, faced significant hurdles in processing massive volumes of data for their fraud detection systems. Their existing data architecture struggled with scalability, data versioning, and the efficient evolution of data schemas, which are critical for developing sophisticated fraud analytics models.

  • Inefficient processing of large data volumes.
  • Lack of efficient data versioning and schema management.
  • High costs associated with their data processing pipeline.

The Solution

Red Analytica architected and deployed a modern data lakehouse on AWS, utilizing Apache Iceberg as the core table format. This solution provided a robust foundation for scalable data management and advanced analytics.

  • Implemented Apache Iceberg for reliable, atomic transactions on the data lake.
  • Enabled efficient schema evolution and data versioning ("time travel").
  • Optimized data processing jobs on AWS, significantly reducing operational costs.

Data Lakehouse with Iceberg

AWS Data Lakehouse with Apache Iceberg

The architecture leverages AWS services with Apache Iceberg to provide a scalable and efficient data platform for Jumio's fraud analytics.

The Impact

25%

Improvement in fraud detection accuracy

40%

Reduction in data processing costs

Time-Travel

Enabled for historical data analysis & model reproducibility