Compliance frameworks, security audits, and quality certifications this company maintains.
Weights & Biases maintains SOC 2 Type II certification for the W&B Cloud platform, demonstrating that its MLOps infrastructure meets rigorous security, availability, and confidentiality controls required by the enterprises and AI labs that store model artifacts, training datasets, and experiment metadata on the platform.
Weights & Biases holds ISO 27001 certification, providing enterprise customers with internationally recognized assurance that the information security management system protecting W&B Cloud environments and customer ML artifacts meets global standards.
Weights & Biases is GDPR compliant, offering data processing agreements and EU data residency options for enterprise customers whose ML training pipelines and model artifacts may include personal data of EU residents.
Weights & Biases complies with the California Consumer Privacy Act, providing enterprise customers with contractual controls needed to operate the W&B platform in environments where ML pipelines process data about California residents.
Weights & Biases supports HIPAA-compliant deployments of W&B through Business Associate Agreements and its Dedicated Cloud offering, enabling pharmaceutical and healthcare AI teams to train and track models on protected health information within a compliant infrastructure boundary.
Security
Penetration Testing
Weights & Biases undergoes annual third-party penetration testing of the W&B Cloud platform, API surfaces, and model artifact storage, with findings remediated and reports made available to enterprise customers under NDA during security due diligence.
W&B Enterprise supports SAML 2.0 single sign-on with Okta, Azure Active Directory, and Google Workspace, allowing ML platform teams to enforce centralized authentication and user provisioning policies across all W&B user accounts.
Security
Role-Based Access Control
Weights & Biases implements role-based access control at the organization, team, and project level, allowing enterprise ML platform teams to restrict which engineers can access specific model artifacts, training runs, and sensitive dataset versions.
Weights & Biases participates in the Cloud Security Alliance STAR Level 1 program, publishing its security self-assessment for the W&B Cloud platform so enterprise ML platform teams can review cloud security controls during vendor evaluation.
Security
Dedicated Cloud Deployment
Weights & Biases offers a Dedicated Cloud deployment option where the W&B platform runs in a single-tenant environment within the customer's own AWS, GCP, or Azure account, meeting data residency and network isolation requirements for regulated industry AI teams.
