Sberbank has opened access to its SberDataScience (SberDS) platform for external business clients. The system supports machine learning and predictive analytics, currently hosting over 20,000 models. Users ranging from non-programmers to data science professionals can utilize the platform for model development.
The platform has been in internal use at Sberbank since 2015, starting with 2,000 models. The bank plans to expand the platform's functionality based on internal experience and feedback from new external users. The product is officially registered in the Unified Register of Russian Software.
SberDS targets three client segments: companies needing a full-cycle development environment, businesses looking to scale model production without increasing headcount, and Sber ecosystem partners. The platform automates routine tasks, allowing business specialists to participate in the development process.
To access the platform, companies must sign a license agreement with the bank. The service is provided via a one-year SaaS subscription, allowing clients to upload their own datasets and build machine learning models within the provided environment.
