Case Study: RareJob Technologies achieves faster ML model development and AI scoring with Amazon Web Services

How RareJob Technologies in education made ML model development more efficient using Amazon SageMaker

RareJob Technologies needed a faster way to build and scale AI-based English speaking test scoring for its PROGOS service, after ML development on local PCs became a bottleneck. The company chose Amazon Web Services, using Amazon SageMaker along with AWS Glue and Amazon Athena to support model development and data access.

Amazon Web Services helped RareJob Technologies move training to a more scalable, lower-cost environment with larger GPU and memory options and managed spot training. As a result, learning time was cut to as little as one-quarter of the previous time, one developer could run multiple training jobs at once, and overall ML development efficiency improved by more than 10x.

RareJob Technologies

Katsuaki Hirakawa

IT Solution Manager