Senior Cloud Infrastructure Engineer - Kubernetes

Responsibilities:

• Foundational hire for AI Ops stack. Own the entire EKS platform: hardened cluster configurations, Terraform modules, Karpenter GPU-aware autoscaling, multi-tenancy (RBAC, namespace isolation, network policies), multi-region DR, and cost optimization.
• Build infrastructure that runs Llama 70B at sub-second latency on multi-GPU instances.
• Design hardened EKS clusters' private endpoints, IMDSv2 pod security admission, image scanning, and audit logging.
• UltraCluster Scale - Experience in building HPCs and large clusters suitable for managing AI ops of SLMs to LLMs.
• Build Terraform modules for the complete Kubogent stack: VPC, EKS, GPU/CPU node groups, IAM, networking, and storage.
• Configure Karpenter for GPU-aware autoscaling across instance families (G6e, P4d, P5 Inferentia).
• Implement multi-tenancy namespace isolation, resource quotas, RBAC, network policies, and fair-share scheduling.
• Build multi-region DR with automated failover, cross-region replication, and failover testing.
• Optimize cloud spend, capacity blocks, spot instances, reserved pricing, right-sizing, and KubeCost integration.
• Design robust network architecture: VPC CNI, private subnets, security groups, Transit Gateway, and private endpoints.

Requirements:

• Experience: 5-9 years.
• 3+ years with EKS/Kubernetes in production.

Must-Have Technical Skills:

• AWS infrastructure: deep VPC, IAM, networking, and multi-account (5+ years).
• Kubernetes/EKS production clusters, networking (CNI), storage, RBAC (3+ years).
• Terraform expert in large module codebases, remote state, workspaces, and CI/CD integration (Karpenter or Cluster Autoscaler) in production.
• GPU instances on AWS G-series (L40S) and P-series (A100), NVIDIA GPU operator/device plugins.
• Security hardening: Pod Security Admission, OPA/Gatekeeper, image scanning, and secrets management.
• Linux systems' performance tuning, storage (EBS, EFS, FSx for Lustre), and kernel parameters.
• Core Tech Stack: Terraform, AWS (EKS, EC2 GPU, VPC, IAM, EBS/EFS/FSx, ECR), Karpenter, Helm, Kustomize, ArgoCD, NVIDIA GPU Operator/DCGM, Calico, Istio, Prometheus/Grafana/KubeCost, OPA/Gatekeeper, Falco, Trivy.