Capability Spotlight: ML Model Management
Establish a single system of record for ML models, integrating AI/ML development into your secure software supply chain.
Overview
More development organizations are building AI/ML models and leveraging them in their software. However, a lack of standard best practices for incorporating MLOps into the software supply chain has led to ML model development largely occurring in isolation. Further, open-source models pose challenges similar to those of other third-party packages (security, availability, versioning, etc.), particularly as their ecosystem is new and threats uncertain. JFrog’s industry-first ML Model Management solution integrates the development and security of AI/ML models with other software components for a unified view. It applies the same best practices organizations use for secure package management to model management, establishing control, availability, visibility, security, and traceability/auditing.
Featured Benefits
Manage All Your Software Artifacts In One Place
Store and manage models alongside the other components that make up modern software applications for better visibility and insight into the status of your software development.
Bring DevOps Best Practices to ML Development
The DevOps practices developed over the past decade, such as artifact management, pipeline automation, and quality/feedback loops can now be applied to ML model management.
Ensure Integrity and Security of ML Models
Manage your models in a system that introduces important controls like RBAC, versioning, license, and security scanning so ML, Security, and DevOps teams feel confident about the models used.
Simplify Model Versioning Across Your SDLC
Leverage custom tags, name and timestamp versioning, and an advanced file system to ensure everyone uses the correct model version—enhancing clarity, context, and scalability.
Key Capabilities
- Secure, advanced AI/ML artifact registry
- Store and manage proprietary, modified OSS and third-party models
- Easy-to-use Python SDK for publishing all model artifacts into Artifactory
- Simplified, intuitive ML versioning
- Proxy Hugging Face, NVIDIA NGC for always available third-party models
- Detect malicious models and enforce license compliance
- Standardize MLOps processes across teams
- Integrated with ML tools such as Jupyter Notebooks, MLFlow, and Amazon Sagemaker
A Single Platform for DevOps, SecOps, and MLOps
The JFrog Platform is the single source of truth for over 7,300 companies. By building and deploying ML models with JFrog, you can seamlessly combine MLOps, DevSecOps, and DevOps workflows into a unified software supply chain platform. Accelerate the delivery of high-quality software while ensuring end-to-end security and compliance across the entire software development lifecycle.
