EKS delivery
Docker Compose on a laptop→repeatable delivery on shared EKS
I owned the delivery path and reviewed and merged three platform cutover changes teammates wrote, covering IAM, ingress, and secrets.
New York, NY · Remote or hybrid · Open to platform roles
Tatia Razmadze
Platform & DevOps Engineer
I have operated AWS and Kubernetes platforms at scale for analytics and healthcare-data teams, with Terraform-managed infrastructure and GitOps delivery.
About
I’ve replaced manual provisioning and deployments with self-service infrastructure and repeatable pipelines. Availability and error-rate alerts give teams a clearer view of what’s happening in production.
That work also includes AI-assisted review workflows I built to surface potential issues in infrastructure changes. Engineers verify the findings and decide what gets merged.
Selected work
Serverless
CDK-generated infrastructure→explicit Terraform, resource by resource
Separate IAM roles per function: the public path can read, never write.
AI tooling
Judgment in someone’s head→checks that run on every change
It caught an RDS instance that never set publicly_accessible. It was also confidently wrong about an IAM policy, and I overrode it — both cases in full →.
Stack
Experience
Platform Engineer
AppsFlyer · 2023—now- 50+ Kubernetes applications moved to Argo CD
- 60+ long-lived AWS keys eliminated through OIDC
- Provisioning reduced from 2–3 days to under 45 minutes
- EKS worker-node spend reduced from $60K to $48K
DevOps Engineer
HealthVerity · 2021—2023- Automated delivery for 12 containerized services
- Operated a pipeline processing two million events daily
Contact
Open to platform and infrastructure roles.