Jumio, an identity verification provider, has developed a real-time feature store on AWS to address challenges like data duplication, manual deployment, and latency in their machine learning operations. The new architecture utilizes services such as Amazon SageMaker Feature Store, Amazon Managed Service for Apache Flink, and Amazon Kinesis Data Streams to enable sub-100ms latency for fraud detection and other real-time predictions. This streaming-first design enhances scalability, reliability, and feature engineering capabilities, streamlining the end-to-end ML development lifecycle for Jumio's fraud detection and digital trust services. AI
IMPACT Enables faster, more accurate fraud detection and identity verification through real-time ML model inference.
RANK_REASON Article details how a specific company (Jumio) implemented an existing technology (AWS services) to solve a business problem, rather than announcing a new product or research.
Read on AWS Machine Learning Blog →
- Amazon Data Firehose
- Amazon EMR
- Amazon Kinesis Data Streams
- Amazon Managed Service for Apache Flink
- Amazon S3
- Amazon SageMaker Feature Store
- Apache Flink
- AWS
- Java
- Jumio
- Python
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
