In the Machine Learning Platform department, we offer solutions to a wide array of challenges across various services by leveraging machine learning. In recent years, as machine learning models have scaled up and become more sophisticated, we have turned to distributed processing using Ray. In this context, we developed Rayleigh as the infrastructure that supports Ray’s operational use, achieving both accelerated development speed and a high level of security.

In this presentation, we will discuss our initiatives, including why we chose Ray and the specific challenges that Rayleigh addresses.