Case Study: RareJob Technologies achieves 10x+ ML development efficiency with AWS SageMaker

RareJob Technologies Uses Amazon SageMaker for English Speaking Tests to Speed Up ML Model Development by More Than 10x

RareJob Technologies, the technical division of RareJob Group, needed a faster, more scalable way to develop machine learning models for its AI-powered English speaking test, PROGOS. The team had been building models locally, which created a development bottleneck and limited how many training jobs developers could run at once.

Using AWS Glue alongside Amazon SageMaker and related AWS services, RareJob Technologies streamlined data acquisition and model development for AI scoring. The company cut ML training time by 25%, improved model development efficiency by more than 10x, and saved about 100 hours of work per month, helping support rapid growth in its speaking test business.

RareJob Technologies

Kentaro Haneda

Executive Officer And Chief Technology Officer