Senior Platform Engineer, AI Infrastructure

Nirvana Labs is building the infrastructure that AI agents and real-time applications run on. The Senior Platform Engineer will be responsible for building the Kubernetes, networking, storage, and automation layers that power the company's Agent Sandboxes and real-time data services.

Responsibilities:

• Operate production Kubernetes for our internal platform and customer workloads, including dataplane, storage, ingress, admission, and the fleet-wide configuration that keeps clusters consistent and self-healing across regions
• Take a cold server from its out-of-band management interface to a fully featured cluster, bootstrapping every layer in between: virtualization, operating system, networking, storage, ingress, and applications. Then move that entire path toward declarative, API-driven provisioning
• Own networking end to end. Give services routable addresses on bare metal without relying on a cloud load balancer, keep tenants isolated at the packet level, and operate DNS and private connectivity across our infrastructure, edge locations, and customer environments
• Lead the next evolution of our networking architecture, including the evaluation and implementation of SDN and overlay technologies
• Build the multi-tenant isolation model that lets customers operate with administrative control inside their own environments on shared infrastructure without ever reaching the underlying platform. That includes admission control, pod security, network isolation, tenancy boundaries, and API-level virtualization
• Keep the platform declarative and reproducible across regions through infrastructure as code, GitOps delivery, automated secrets and certificate management, pull-through registries, and automated TLS. Standing up a new region should be routine, not heroic
• Operate the applications that make the platform useful, including the registry, secrets systems, GitOps tooling, and management plane, as first-class production services that are just as important as the infrastructure beneath them
• Bring real disaster recovery instincts to the systems behind those services: resilient storage, tested snapshots, rehearsed restores, and clear recovery procedures
• Stand up and own platform observability across metrics, logs, traces, and flow-level visibility into the dataplane
• Connect that visibility to alerting that surfaces meaningful failure modes before customers feel them, using open and vendor-neutral instrumentation wherever practical
• Harden the platform across cloud, Kubernetes, operating system, storage, and network layers through audit trails, encryption at rest, sensible baselines, access controls, and secure defaults. Build compliance into the platform from the beginning rather than treating it as a layer added afterward
• Share the on-call rotation, lead incident response when necessary, and keep runbooks and design documents sharp. The next person debugging a problem at 3 a.m. should be able to find what they need

Requirements:

• 5+ years of experience in infrastructure, SRE, systems engineering, or platform engineering, including significant experience running Kubernetes in production
• Depth and adaptability matter more to us than experience with any one distribution. RKE2, kubeadm, upstream Kubernetes, k3s, Talos, and similar platforms are all relevant
• Deep networking expertise, which we weight heavily. You should understand the CNI and network-policy layers, load balancing, DNS, L2 and L3 networking, BGP, ingress, Gateway API, and how an eBPF dataplane such as Cilium actually moves packets
• You'll help evaluate SDN and overlay options, including technologies such as Kube-OVN, and help stand up what we choose
• A real understanding of Kubernetes internals that goes beyond driving kubectl: scheduling, controllers, reconciliation, the admission chain, CRDs, and the CNI and CSI ecosystems
• Deep Linux administration and troubleshooting experience across Ubuntu or RHEL-family systems, including systemd, the network stack, kernel behavior, sysctl tuning, and packet capture analysis
• Infrastructure as code and GitOps as daily practice, including Terraform or Terragrunt and declarative delivery through Argo CD or equivalent systems
• Fluency in Git-based review workflows and CI systems such as GitHub Actions
• Programming ability for real tooling and automation using Go, Python, Bash, or similar languages
• Production experience on at least one major cloud platform such as AWS, GCP, or Azure
• Sound instincts for stateful systems and disaster recovery
• A solid security foundation, including Kubernetes hardening, admission control, network isolation, secrets management, identity, permissions, and IAM
• Familiarity with cloud-native observability, including metrics, logs, traces, Prometheus, Grafana, Loki, OpenTelemetry, or equivalent systems
• Comfort working in AI-accelerated engineering workflows
• Clear technical communication
• Fluent English and the ability to maintain meaningful working-hours overlap with the US team
• Bare-metal provisioning and lifecycle automation experience, including BMC, IPMI, Redfish, PXE, image-based installation, Cluster API, Metal3, or similar systems
• Experience with immutable, API-driven node operating systems such as Talos
• VM-on-Kubernetes virtualization experience, including KubeVirt or related technologies
• Hands-on CSI storage integration and experience operating stateful workloads on Kubernetes
• Experience building multi-tenant infrastructure and platform isolation using tools and approaches such as admission control, Kyverno, OPA, Pod Security, network policies, virtual clusters, or vCluster
• Experience with test-driven infrastructure and tooling such as Kyverno Chainsaw, Terratest, policy testing, infrastructure validation, or similar approaches
• Strong opinions about how to use coding agents effectively on large infrastructure codebases
• Depth in a specific datastore such as PostgreSQL or ClickHouse
• Experience with ultra-low-latency or high-throughput networking
• Prior SRE or reliability engineering experience
• Comfort working directly with customers when the situation calls for it
• You write or speak publicly about infrastructure, distributed systems, Kubernetes, networking, SRE, databases, or related technical topics