Staff+ Software Engineer, Capacity Engineering

About the role

Anthropic is building reliable, interpretable, and steerable AI systems. The Capacity Engineering team ensures that the company’s massive and rapidly growing compute fleet is accurately accounted for, efficiently utilized, and properly allocated. As a Staff+ Software Engineer you will design and operate the data and tooling that give the organization real‑time visibility into fleet health and utilization.

Key responsibilities

• Build and maintain data pipelines that ingest and normalize telemetry from heterogeneous cloud and accelerator environments.
• Develop observability tooling and dashboards that provide real‑time insight into cluster health, capacity planning, and allocation issues.
• Instrument performance metrics across training, inference, and evaluation workloads to measure hardware efficiency.
• Own the production systems that serve BigQuery tables used by research, finance, and leadership teams.
• Collaborate with research engineering, infrastructure, inference, and finance groups to align capacity decisions with business priorities.

Required profile

• Comfortable moving between data engineering, systems engineering, and observability tasks.
• Thrives in high‑autonomy, high‑ambiguity environments and can make decisions that impact large spend areas.
• Able to write production‑quality code daily and operate at scale on Kubernetes‑native infrastructure.
• Strong communication skills to work closely with cross‑functional teams.

Required skills

• Kubernetes (deployment, operations, scaling)
• Data pipeline development and telemetry ingestion
• Observability tooling and dashboard creation
• Performance instrumentation for hardware utilization
• Experience with cloud platforms and multi‑accelerator environments
• BigQuery for data storage and querying