Senior Devops Engineer

About ParcleParcle is building a cloud-deployed AI agent platform on top of open-source agent frameworks. Our mission: when a user signs up, their agent should be ready in seconds (not minutes). We're building the infrastructure to make that happen, including runtime provisioning, warm-pool allocation, tenant bootstrap, and the operational systems that keep it fast, safe, and cost-efficient.Role DescriptionWe're looking for a Senior DevOps Engineer to own the platform that provisions, assigns, and operates agent runtimes at scale. This isn't a generic DevOps role — you'll be building and operating a control plane for fast per-user runtime allocation with strong reliability, tenant isolation, and cost discipline.You'll own the full provisioning lifecycle: warm instance pools, signup-to-ready pipelines, tenant bootstrap automation, cold-start fallbacks, capacity forecasting, multi-tenant isolation, observability, and cost controls.What Success Looks Like (First 90 Days)p95 signup-to-agent-ready under 30s. Warm-pool hit rate above 90%. Provisioning reliability at 99.9%. Clear dashboards, alerts, and runbooks in place.QualificationsMust have:6+ years in Platform Engineering, DevOps, SRE, or infrastructure engineeringExperience building low-latency provisioning or orchestration systems in productionDeep Kubernetes, Terraform, and AWS (ideally EKS) experienceTrack record operating multi-tenant systems with real isolation constraintsStrong autoscaling, capacity planning, and incident management skillsHands-on security fundamentals: IAM, secrets management, network controls, audit loggingNice to have:Workload-level cost optimization experienceExperience with sandboxed runtimes, agent infrastructure, or container poolingFamiliarity with queues, streaming systems, Postgres, Redis, OAuth/OIDC patternsTech StackKubernetes · Terraform · AWS · GitOps (ArgoCD/Flux) · CI/CD · Prometheus · Grafana · OpenTelemetry · KMS/SSM/VaultWhy This RoleIf you've built provisioning systems, control-plane workflows, or low-latency multi-tenant platform services in production — and you want to do it for AI agents at scale — this will feel like home.