IAM and Privileged Access Foundations for Secure ML Platforms
summary
Describes an implemented identity stack that ties HR as the authoritative source to automated provisioning, role changes, deprovisioning and rehire workflows, integrates PAM features like credential vaulting and automated rotation, and adds session brokering, activity monitoring and just‑in‑time elevation for service and admin accounts. Claims unified IAM+PAM gives a single audit trail and recurring certification campaigns to catch privilege creep, supports HIPAA SOC GDPR audits, and enforces least‑privilege across ML pipelines and CI/CD deployments.
AI-written summary · about
chapters (10)
- 0:00Welcome & What This Talk Covers (IAM + PAM for ML Platforms)
- 0:58Modern IAM Challenges in Enterprise Machine Learning
- 2:04Identity Lifecycle Management Basics: Provisioning to Deprovisioning
- 2:56Lifecycle Events & HR as the Authoritative Source (Onboard/Move/Offboard/Rehire)
- 4:10Access Governance: Certification Campaigns & Least Privilege
- 4:57Role-Based Access Control (RBAC) for ML Teams: Who Needs What
- 6:33Privileged Access Management (PAM): Securing Admin & Service Accounts
- 7:18PAM for ML Pipelines: Vaulting, Rotation, Session Monitoring
- 9:06Unifying IAM + PAM for Centralized Governance & Compliance
- 10:58Key Takeaways & Closing Thanks
related talks
35mIdentity and Access Management for the Decade Ahead
31mIAM and why it is important in Cloud environments
05mScaling Machine Learning Without Compromising Privacy
01h 02mWhy IAM is the new perimeter, and how to govern it
50mAd Hoc Mutable Infrastructure for Security Management
28mAn Approach to Protect the Entire Privileged Access Life Cycle
16mIdentity-Centric Security for AI Systems
19mSecuring AI Pipelines from Development to Production