Sandeep Chakka

WORK HISTORY

EDUCATION

ABOUT SANDEEP CHAKKA

Around 14 years of IT experience designing, implementing, and optimizing cloud infrastructure solutions, with deep expertise in DevOps culture, CI/CD pipeline engineering, build/release management, and cloud-native architecture on AWS.Built custom AMIs from scratch tailored for online and mobile banking workloads, serving as the foundational base images for provisioning plain vanilla Kubernetes clusters on AWS — ensuring standardized, hardened, and repeatable node configurations across dev, staging, and production environments.Led the Kubernetes onboarding of Personetics integration single-handedly, acting as the primary K8s engineer responsible for platform configuration, workload deployment, networking, and ensuring seamless integration within the existing Kubernetes ecosystem.Deployed and managed GPU workloads on Kubernetes using node selectors, taints/tolerations, and resource quotas to ensure efficient GPU allocation across ML training and inference pipelines — configuring Cluster Autoscaler and Karpenter on EKS to dynamically provision and deprovision GPU nodes based on workload demand.Implemented Kubernetes RBAC policies, namespaced service accounts, and network policies to enforce security boundaries between workloads, teams, and GPU job queues — alongside IRSA (IAM Roles for Service Accounts) for fine-grained pod-level AWS permissions without static credentials.Designed and built custom Docker images for critical shared services infrastructure components including HashiCorp Vault, Consul, Jenkins, and Nexus — optimized for project-specific requirements and deployed within a shared services model to support multiple application teams across the platform.Automated infrastructure provisioning using Terraform and AWS CloudFormation, managing EKS clusters, node groups, IAM roles, networking components, and GPU instance launch templates as code — leveraging remote state management, workspaces, and modules across multi-environment (dev/staging/prod) deployments with full auditability.Authored reusable Terraform modules for Datadog monitors, Cloud Custodian policies, and EKS infrastructure — reducing provisioning time, eliminating configuration drift, and enabling consistent governance across multiple application teams.Architected and managed AWS environments across EC2, EKS, VPC, IAM, S3, RDS, and GPU-optimized instance families (P3, P4, G4, G5) — designing scalable VPC architectures encompassing multi-AZ public/private subnets, Internet Gateways, NAT Gateways, VPC peering, and VPC endpoints for banking-grade workloads.