AWS · Kubernetes · DevOps · AI · Automation. Engineering resilient, cost-optimized cloud platforms — from enterprise Landing Zones to Kubernetes-native DevOps and AI-driven automation.
I'm Pankaj Kumar, a Senior AWS Cloud Solutions Architect with 15+ years of IT experience, currently architecting enterprise-scale cloud platforms. My work spans AWS Landing Zone design, large-scale VM migrations, Kubernetes/EKS-based DevOps platforms, and increasingly, AI-driven automation and cloud cost governance.
I mentor a team of 3 cloud engineers and operate a personal multi-account AWS Organizations lab environment where I prototype and validate architecture patterns before they reach production engagements.
Core areas: AWS Landing Zone & Multi-Account Governance · Enterprise Cloud Migration · Kubernetes/EKS & GitOps CI/CD · Cloud Cost Optimization · AI/ML-Integrated Architectures
Enterprise engagements and personal lab builds. Client details are anonymized; architectures and outcomes are real.
Client needed a secure, governed multi-account AWS foundation to migrate production VMs off legacy VMware infrastructure, with strict RTO/RPO requirements.
Designed a Landing Zone across management, network-hub, workload-prod, log-archive, and audit accounts. Implemented Transit Gateway-based hub-spoke networking, centralized logging, and guardrails via AWS Organizations SCPs.
Client needed a modern CI/CD pipeline and container orchestration platform to replace manual deployment processes.
Built path-scoped CI/CD pipelines deploying to EKS via GitOps patterns for automated, auditable releases.
Personal multi-account AWS lab was accumulating orphaned resource costs — idle Transit Gateway attachments, unused Elastic IPs, an unused VPC Interface Endpoint — with no automated detection.
Built a serverless cost governance pipeline: Lambda scans for orphaned resources, DynamoDB tracks state, SNS alerts on findings, API Gateway exposes a dashboard endpoint. Fully deployed via Terraform.
REST API via API Gateway, Lambda backend, DynamoDB storage, EventBridge for async workflows, Cognito for authentication, plus a private API endpoint pattern.
Deployed a MongoDB replica set via Helm charts, integrated with Secrets Manager through IRSA, configured PodDisruptionBudgets and HPA for resilience/scaling, monitored via kube-prometheus-stack.
Hands-on infrastructure code, open for review.