SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 405,000 - 485,000 / annual
Anthropic operates one of the industry's largest AI compute fleets spanning multiple cloud providers and datacenters to train, research, and serve frontier AI models. The Kubernetes Platform team owns the control plane that powers these fleets at unprecedented scale.
In this role, you will own and extend Anthropic's Kubernetes infrastructure to support AI workloads at scale where standard defaults no longer apply. Key responsibilities include:
- Own, operate, and extend the Kubernetes scheduler for accelerator fleets, including custom scheduling plugins for gang scheduling, topology awareness, and preemption
- Scale the Kubernetes control plane (apiserver, etcd, controller-manager) to support clusters far beyond typical limits, proactively identifying bottlenecks
- Design, build, and operate core cluster services like service discovery that every workload depends on
- Build 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 required features and escalations
- Participate in on-call rotations, lead incident response, and design reliability processes (postmortems, runbooks, SLOs)
You will need significant production distributed systems experience, deep hands-on Kubernetes expertise (scheduler, controllers, apiserver, large multi-tenant clusters), proficiency in systems languages (Go, Python, Rust, C++), and demonstrated ability to debug complex issues across the full stack. Strong communication and consensus-building skills are essential.
Preferred qualifications include Kubernetes internals contributions, experience with cluster schedulers (Kueue, Volcano, Slurm), control plane scaling, ML infrastructure familiarity (GPUs, TPUs, gang scheduling, NCCL), GCP/AWS experience, and 12+ years in relevant infrastructure roles.
