Staff Software Engineer, Kubernetes Platform

About the role

Anthropic is building one of the industry’s largest AI compute fleets that run on Kubernetes across multiple clouds and datacenters. The Kubernetes Platform team owns the control plane that schedules and operates thousands of accelerator nodes, ensuring reliable training of frontier AI models. As a Staff Software Engineer you will design, scale, and operate core platform components that keep the fleet fast, correct, and always available.

Key responsibilities

• Own, operate, and extend the Kubernetes scheduler with custom plugins for gang scheduling, topology awareness, and preemption.
• Scale the control‑plane components (apiserver, etcd, controller‑manager) to support clusters far beyond typical limits.
• Design, build, and run core services such as service discovery that all workloads depend on.
• Develop and maintain custom controllers, operators, and CRDs.
• Partner with research, training, and inference teams to translate workload requirements into platform capabilities.
• Collaborate with cloud providers on feature development and escalation handling.
• Participate in on‑call rotation, lead incident response, and create post‑mortems, runbooks, and SLOs.

Required profile

• Significant experience building and operating production distributed systems.
• Deep, hands‑on expertise with Kubernetes internals (scheduler, controllers, apiserver, large multi‑tenant clusters).
• Proficiency in at least one systems language such as Go, Python, Rust, or C++.
• Strong debugging skills across API, node, and network layers.
• Track record of designing for reliability, correctness, and clear failure semantics.
• Excellent written and verbal communication and ability to build consensus.

Required skills

• Go, Python, Rust, C++
• Kubernetes (scheduler, controller‑runtime, client‑go, apiserver, etcd)
• Custom scheduling frameworks (gang scheduling, topology‑aware placement)
• Cloud platforms: GCP, AWS (including GKE/EKS)
• Infrastructure as Code
• Linux kernel tuning, cgroups, eBPF
• Accelerator hardware: GPUs, TPUs, Trainium, NCCL
• Batch systems: Kueue, Volcano, Slurm

What we offer

• Competitive salary £325,000–£485,000 GBP per year.
• Equity donation matching, generous vacation and parental leave.
• Flexible working hours and hybrid office policy (minimum 25% in‑office).
• Access to cutting‑edge AI research and large‑scale compute infrastructure.