Principal Engineer, Capacity/Compute | Anthropic Careers

About this role

Principal Engineer, Capacity/Compute | Anthropic Careers

Anthropic

Hybrid · San Francisco, California Senior Full-time USD 320k–485k / yr Posted 5 days ago International

Principal Engineer, Capacity/Compute | Anthropic CareersAnthropic

Apply on company site ↗

Not ready to apply yet?

About the role

Anthropic runs one of the largest and fastest-growing infrastructure fleets in the industry, spanning multiple accelerator families, CPU families, and clouds. The Capacity Engineering team makes sure every infrastructure resource is accounted for, well utilized, and efficiently allocated. You will build the production systems behind this work: data pipelines that ingest and normalize telemetry from different cloud environments, observability tooling for real-time fleet visibility, and performance instrumentation that measures how efficiently workloads use their hardware. You will work with research engineering, infrastructure, inference, and finance teams on decisions around one of Anthropic’s largest areas of spend.

This is a pipeline role feeding four focus areas, and you will concentrate mainly on one:

• Data platform: pipelines that ingest utilization telemetry and normalize billing across cloud providers into BigQuery
• Planning: real-time fleet visibility, cluster health tooling, capacity planning, and alerting
• Efficiency: measuring and improving how well each major workload uses its hardware
• Attribution and forecasting: connecting fleet cost to business activity and building a defensible compute plan

Key Responsibilities

• Build the planning and allocation stack, including cross-region and cross-provider placement, guardrails, queueing, and occupancy KPIs
• Drive efficiency programs such as stranding and rightsizing, unused capacity recovery, and job-level utilization across training, inference, and eval
• Own attribution and forecasting by reconciling billing across ten-plus providers against telemetry and internal systems, then turning demand signals and research roadmaps into a compute plan
• Build the data platform with ownership of completeness, latency SLOs, and gap detection, including integrating each new provider
• Operate Kubernetes-native systems at scale, including collection agents, workload labeling, and taint, reservation, and scheduling behavior
• Treat outputs as a product by gathering requirements, defining schema contracts, and designing for consumers from research engineers to a CFO, including on-call and SLOs

Requirements

• Strong track record building and operating production systems (hands-on engineering with a devops flavor)
• Production-quality Python and SQL, including BigQuery SQL, table-valued functions, and views
• Deep experience with at least one major cloud provider (AWS, Google Cloud, or Azure) and its operations
• Experience with Prometheus, PromQL, and Grafana, including writing recording rules and building monitoring that engineering teams rely on
• Ability to gather your own requirements and work across organizational boundaries in an ambiguous environment
• Bachelor’s degree or an equivalent combination of education, training, and experience

Skills

Technical Skills

• Python and SQL (BigQuery)
• Kubernetes
• Prometheus, PromQL, Grafana
• Cloud platforms (AWS, GCP, or Azure)
• Data pipelines, schema contracts, and SLOs
• Capacity planning, cost attribution, and forecasting

Soft Skills

• Autonomy and comfort with ambiguity
• Cross-team collaboration (research, infrastructure, inference, finance)
• Requirements gathering and product thinking
• Strong communication

Nice to Have

• Capacity planning, resource management, or cost attribution experience at a hyperscaler or large-scale ML environment
• Scheduling and packing efficiency experience, or profiling-driven optimization of large distributed workloads
• Multi-cloud data ingestion experience (billing exports, reservation APIs, capacity reservations, commitments, vendor telemetry)
• Total cost of ownership and forecasting experience
• Accelerator infrastructure familiarity (GPU metrics such as DCGM, TPU utilization, Trainium metrics)
• Experience building internal data products with self-service access, schema contracts, API serving, and documentation
• Storage efficiency, retention, and lifecycle program experience at exabyte scale

Job Details

• Location: US (company headquartered in San Francisco)
• Work Mode: Hybrid, with staff expected in an office at least 25% of the time (some roles may require more)
• Salary: $320,000 - $485,000 USD annually
• Visa Sponsorship: Available (not guaranteed for every role and candidate)

Benefits

• Competitive compensation and benefits
• Optional equity donation matching
• Generous vacation and parental leave