Selected Work
Selected DevOps and AI pipeline work — real systems, in production.
Faster ticket triage
Click a step to see what it does.
40% lower node cost
Migrated the platform to EKS with Karpenter autoscaling and OpenTofu-managed infrastructure: cut node costs 40% via ARM64 and spot workloads, removed manual provisioning.
Self-serve, auditable releases
Replace manual deploys with GitOps (ArgoCD + Traefik) across 12+ microservices: teams ship and audit their own releases, no single-person bottleneck.
Downtime cut to seconds
Use ArgoCD self-healing rollouts so a replacement pod is confirmed healthy before its predecessor is ever taken down, cutting website downtime to seconds per month.
Encrypted at rest
Hardened secrets management with SOPS/KMS across services, removing plaintext credentials from repos and deploy pipelines.
One CI/CD pattern, not many
Standardized CI/CD on CodeBuild across services, leading to fewer deploy-time inconsistencies.
Issues caught before users notice
Set up metrics, logs, and alerting across all services, surfacing incidents before customers report them instead of finding out after the fact.