Work

Building and operating Crusoe's next-generation Managed Services platform for AI and HPC workloads. The platform supports Kubernetes and Slurm-based workloads at scale, designed for the operational demands of GPU-intensive AI infrastructure.

Driving infrastructure automation, operational excellence, and the cross-functional alignment needed to deliver reliable, cost-efficient compute to Crusoe's AI customers.

Led the design and implementation of Workday's cloud-native infrastructure for its next-generation application platform — shipping two applications to production with full end-to-end infrastructure ownership.

Designed and built Workday's data and observability platform: an EKS-based telemetry system ingesting 30+ TB daily across all Workday environments in private and public clouds, centralizing the telemetry collection that had previously been fragmented across dozens of systems. The platform significantly reduced capital expenditure while improving system-wide monitoring and analytics.

Architected and led the migration of a 10PB operational DataLake from private data centers to AWS — a two-phase approach that decoupled raw data transfer from transformation, enabling thorough validation and reuse of business logic in the downstream streaming pipeline. Delivered without disrupting production workloads.

Engineered and managed the Workday Kubernetes platform for public cloud delivery, building Zero Down Time deployment capabilities and working directly with service teams to migrate their workloads to AWS and Kubernetes.

Led the migration from OpenStack to AWS for Symantec Endpoint products. Spearheaded Cloud Governance initiatives that reduced infrastructure costs by 8%. Implemented containerization of microservices using Docker and Kubernetes.

Established in-house DevOps best practices and built self-service CI/CD pipelines for microservices and infrastructure automation. Led the production deployment of Symantec's first cloud-native product — the company's first step into the public cloud.

Developed a data curation and management platform for seismic, ground, and satellite datasets at USC's Semantic Information Research Lab, in collaboration with NASA JPL. Before this work, earthquake research data was scattered across incompatible formats and isolated research papers — there was no shared repository.

The platform unified heterogeneous datasets under common science standards and exposed them through a data processing and delivery API integrated with GIS and mapping tools. It was subsequently used for earthquake early-warning research, the California ShakeOut preparedness exercise, insurance risk modeling, and ongoing seismic science.