A decade of experience in DevOps, microservices and container orchestration. These days I spend most of my time on Amazon EKS — making clusters cheaper to run, easier to upgrade, and capable of hosting GPU and generative AI workloads without the operations team paying for it later.
- 🏗️ Container Architect & GenAI Practitioner at Amazon Web Services
- 🔬 Working on multi-tenant GPU platforms, agentic AI on Kubernetes, and EKS day-2 operations
- ✍️ Published on the AWS Containers Blog and AWS Builder Center
- 🌱 Open source contributor — AWS EKS Workshop, app modernization blueprints, Istio on EKS
- 📍 Based in the New York metro area
Clouds — AWS (primary), GCP, Azure
Container & Orchestration: Amazon EKS, EKS Auto Mode,EKS capabilities, Karpenter,Helm, Argo CD,Flux
Networking: VPC CNI ,Cilium, Gateway API, kgateway, CoreDNS
Provisioning & Platform Delivery — Terraform (+ pre-commit-terraform, tflint, Checkov), Crossplane & Upbound reference platforms, AWS CDK, CloudFormation, Ansible · internal developer platforms and self-service APIs
Observability: OpenTelemetry,Prometheus, Grafana, AMP/AMG + ADOT, Langfuse, Hubble
GitOps & Pipelines — Argo CD, Argo Rollouts, Argo Workflows, Flux · multi-cluster GitOps topologies · GitHub Actions, Jenkins, AWS CodePipeline / CodeBuild
Traffic, Mesh & Connectivity — Istio, Envoy (external processing filters) · Gateway API, AWS Load Balancer Controller, ALB/NLB · VPC CNI, CoreDNS, ExternalDNS, cert-manager,Route 53, CloudFront, PrivateLink, Transit Gateway
Accelerated Compute & Training — NVIDIA GPU Operator, device plugin, DCGM exporter, MIG partitioning · Karpenter GPU provisioning, Spot capacity, EFA networking · multi-tenant scheduling, quotas and cost attribution on EKS
AI/ML Inference, Serving & Gateway: vLLM,NeMo Framework, AWS Neuron / Trainium2 / Inferential2, Ray Serve, NVIDIA Triton, Amazon Bedrock AgentCore,kgateway + agentgateway, Bifrost, LiteLLM
AI Agents Frameworks : Strands Kagent, MCP, LangGraph, CrewAI, langchain
Security, Identity & Compliance: IRSA, EKS Pod Identity,External Secrets Operator, AWS Secrets Manager, Prowler, Kyverno, OPA Gatekeeper, Trivy · GuardDuty, Inspector, Security Hub, Keycloak, OAuth2-Proxy (OIDC)
MLOps: MLflow, Kubeflow, KServe
Vector DB & Evaluation: Milvus, RAGAS, LLM-as-a-Judge, Langfuse Custom Evaluators
Languages: — Go, Python, Bash
NVIDIA Stack: GPU Operator, Network Operator, k8s-device-plugin, DCGM / DCGM Exporter · MIG partitioning, time-slicing, KAI Scheduler (Run:ai) · CUDA, cuDNN, NCCL, NVLink / NVSwitch, GPUDirect RDMA · TensorRT, TensorRT-LLM, Triton Inference Server, NIM microservices, NeMo Framework & Guardrails, Dynamo · DGX / HGX (H100, H200, GB200), BlueField DPU & DOCA · NGC catalog, AI Enterprise, Nsight profiling
Migration Strategy — 7 Rs assessment (rehost, replatform, repurchase, refactor, retire, retain, relocate) · portfolio discovery and wave planning · Migration Acceleration Program (MAP) engagements · TCO and Migration Evaluator business cases
Kubernetes & Container Migrations — self-managed Kubernetes → Amazon EKS · ECS → EKS · OpenShift → ROSA · on-prem → EKS Anywhere and Hybrid Nodes · Velero backup/restore and cluster relocation · in-place EKS version upgrades with eksup · CNI swaps (VPC CNI ↔ Cilium), Ingress → Gateway API, Cluster Autoscaler → Karpenter
Application Modernization — monolith decomposition and strangler-fig migrations · containerization and 12-factor refactoring · serverless with Lambda, Fargate and Step Functions · event-driven architecture with EventBridge, SQS/SNS and MSK · API-first with API Gateway and gRPC
- Full request and response compliance logging on Amazon EKS · AWS Containers Blog
- Better Together: Amazon EKS Auto Mode and Istio Ambient Mesh · AWS Containers Blog
- Enabling swap memory on EKS with Karpenter · AWS Builder Center
- Optimizing IP usage with subnet selection in EKS Auto Mode · AWS Builder Center
- Automating Kubernetes version monitoring and upgrades · AWS Builder Center
More on AWS Builder Center