Job Description:ย
Data and AI platform company building managed data products for federal and enterprise clients. Our stack is multi-cloud - primarily Google Cloud Platform (GKE, BigQuery, Cloud Monitoring) with AWS workloads and cross-cloud networking - orchestrated via Terraform, Kubernetes, and CI/CD (GitHub Actions). We're hardening platform reliability and operations as part of a broader push toward SOC 2 compliance.
ย
ย
The Opportunity
We're scaling a multi-cloud, multi-cluster Kubernetes platform that runs data ingestion, orchestration (Airflow), search (Vespa), and API workloads across Google Cloud Platform and AWS. Much of the infrastructure exists as Terraform and Helm, but reliability, observability, and operational maturity still need ownership - from networking and node-pool design through incident response and CI/CD hardening.
ย
You will own the reliability and operability of our Kubernetes and cloud platform: keep clusters healthy across clouds, make deployments safe and repeatable, and turn tribal ops knowledge into runbooks and alerts. This is a hands-on build role with strong mentorship and direct exposure to architecture decisions.
ย
ย
What You'll Do
- Own day-to-day reliability of Kubernetes clusters (GKE / EKS): node pools (Spot/standard), autoscaling, taints/labels, and workload placement.
- Maintain and extend Terraform modules for VPC, clusters, IAM, and add-ons across Google Cloud Platform and AWS; land all infra changes through Git and PRs.
- Design and operate cross-cloud networking and connectivity (VPC, peering, VPN/Interconnect, DNS, private service access).
- Operate and improve Airflow-on-Kubernetes (executor configs, resource requests, pod scheduling, and failure recovery).
- Build and harden CI/CD (GitHub Actions) for infra and app deployments - gated promotes, environment isolation, rollback paths.
- Design observability: metrics, logs, traces, SLOs/SLIs, and actionable alerting across clouds (not alert noise).
- Write runbooks for common failure modes; lead or support incident response and postmortems.
- Harden security posture for SOC 2: least-privilege IAM, secret management, audit logging, change control.
- Partner with data and platform engineers on capacity, cost, and performance trade-offs.
- Participate in design discussions, surface trade-offs, and propose improvements as you encounter them in the stack.
ย
ย
Required Qualifications
- 4โ5 years in DevOps, SRE, or platform engineering.
- Strong hands-on Kubernetes experience in production (GKE and/or EKS) - scheduling, resources, networking, troubleshooting.
- Multi-cloud experience - production work on both Google Cloud Platform and AWS (or deep in one with solid working knowledge of the other).
- Solid Terraform experience - modules, state, reviews, and safe apply workflows you have used in production.
- Working knowledge of CI/CD (GitHub Actions, Cloud Build, CodePipeline, or similar) for infra and containerized apps.
- Comfortable with Linux, containers, and debugging production issues from logs/metrics.
- Cloud fundamentals across providers: IAM, VPC/networking, managed Kubernetes, secrets, monitoring/logging.
- Clear written communication - runbooks, PR descriptions, incident notes.
ย
ย
Preferred Qualifications
- Airflow on Kubernetes / GKE operations.
- Helm / Kustomize for app and platform packaging.
- Prometheus, Grafana, Cloud Monitoring, and/or CloudWatch at scale.
- Experience with Spot/preemptible node pools and cost-aware capacity planning across clouds.
- Prior work in SOC 2 or similarly controlled environments.
- Exposure to data platforms (BigQuery, dbt, ingestion pipelines) as a platform consumer.
ย
ย
What You Bring to the Team
A hands-on operator who has kept production Kubernetes platforms healthy under real load across more than one cloud - comfortable tracing a failed deploy or flapping pod back to root cause, and turning that into better automation, alerts, and docs. Someone who defaults to Infrastructure-as-Code, treats alerts as a product, and asks good questions in design discussions.
ย
