Sheshank Dudaboina

Operate production GPU inference infrastructure across 70+ Kubernetes clusters and roughly 20 cloud providers. Build clusters and their model-serving, networking, storage, autoscaling, GitOps, and telemetry stacks; lead P0 response and vendor escalations across large NVIDIA H100, B200, and A100 fleets.

Founded the platform SRE function and owned reliability, observability, cost efficiency, and production operations across multi-region AWS and EKS. Built the centralized VictoriaMetrics, Grafana, Loki, and Alertmanager platform and drove resource right-sizing that saved about $15K per month.

Contributed to planning a payment-platform migration from an on-premises data center to AWS, including secure and highly available architecture plus reusable Terraform and Terragrunt templates.

Operated large-scale AWS production environments supporting global PlayStation services, including the PS5 launch. Led a Datadog migration, moved legacy workloads to EKS, automated infrastructure, and built safer CI/CD pipelines that reduced deployment time by about 40%.

Designed AWS infrastructure and multi-region deployments for high-throughput financial systems, built Kubernetes platforms with KOPS and GitLab CI/CD, and improved observability for real-time transaction workloads.

Implemented cloud automation across AWS and OpenStack, deployed highly available architectures, and automated infrastructure and configuration with Terraform and Ansible.