JFrog, AWS team up for machine learning in the cloud

Software supply chain provider JFrog is integrating with the Amazon SageMaker cloud-based machine learning platform to incorporate machine learning models into the software development lifecycle.

The JFrog platform integration with Amazon SageMaker, available now, ensures artifacts produced by data scientists to develop machine learning applications are pulled from and saved in the JFrog Artifactory. Announced on January 17, the JFrog-Amazon pairing makes machine learning models immutable, traceable, secure, and validated as they mature for release, JFrog said.

The company also unveiled versioning capabilities for its ML Model Management platform, intended to help ensure that compliance and security are incorporated at each step of model development.

The JFrog Amazon SageMaker integration applies DevSecOps practices to machine learning model management, so developers and data scientists can expand and secure the development of machine learning projects in an enterprise-grade manner, JFrog said. The platform integration with SageMaker is available for current users. By working with AWS, JFrog has designed a workflow indoctrinating these practices to machine learning model development in the cloud, providing speed and security.  Users of the integration can anticipate the following benefits:

• Bring machine learning closer to software development and production lifecycle workflows, protecting models from modification or deletion.
• Development, training, and securing of machine learning models.
• Detecting and blocking the use of malicious models.
• Scanning of model licenses to ensure compliance with regulatory requirements and company policies.
• Blocking the use of malicious models across an organization.
• Distributing machine learning models as part of a software release.

JFrog also has added new capabilities for its ML Model Management platform, bringing model development into an organization’s secure SDLC. Versioning capabilities increase transparency around model versions. Projects are to abide by regulatory and organizational compliance.