Context
Data scientists and analytical teams need governed access to cloud notebooks to develop and run their workloads. Provisioning must support enterprise scale while maintaining access controls and workload isolation.
My role
As Principal Product Manager at American Express, I owned the product capabilities for Terraform-based notebook and Vertex AI Workbench provisioning, including API contracts, user experience, and non-functional requirements.
Solution
Delivered self-service, on-demand notebook provisioning through Lumi Portal and API-driven services, with Terraform orchestrating infrastructure provisioning on GCP.
- Self-service access: A portal experience for notebook users alongside headless API-driven capabilities.
- Infrastructure orchestration: Terraform-based provisioning for notebooks and Vertex AI Workbench.
- Scalable shared infrastructure: Multi-tenant capabilities with quotas and workload isolation.
- Enterprise controls: Requirements for role-based access, least privilege, auditability, provisioning latency, recovery, monitoring, and cost controls.
Impact
Lumi Portal served 3,000+ notebook users within a broader cloud migration serving 10,000+ users. The product made governed notebook provisioning available through reusable platform capabilities.
Technologies & capabilities
GCP, Vertex AI Workbench, Terraform, API/service contracts, multi-tenant infrastructure, and governed access.