Principal Engineer Capacity/Compute

Anthropic•Seattle, WA

5d•$320,000 - $485,000•Hybrid

About The Position

Anthropic’s Capacity Engineering team ensures its infrastructure resources are accounted for, well utilized, and efficiently allocated across multiple accelerator and CPU families and cloud providers. The team owns the data, tooling, and operational systems used to plan, measure, and maximize utilization across first-party and third-party compute. The engineer will build and operate production systems for capacity engineering, including telemetry data pipelines, fleet observability tooling, and performance instrumentation for workloads. The role spans data engineering, systems engineering, and observability, and involves Kubernetes-native infrastructure at scale. The engineer will collaborate with research engineering, infrastructure, inference, and finance teams and contribute to decisions involving substantial infrastructure spending. This is a pipeline role across four overlapping areas: data platform, planning, efficiency, and attribution and forecasting. Depending on background and business priorities, the engineer will focus primarily in one area. The work includes delivering reliable, timely, discoverable data; making fleet state actionable; measuring and improving workload efficiency; and connecting infrastructure costs and demand signals to business and research plans. The role is hands-on and production-focused, with a DevOps flavor. Anthropic currently expects staff to work from one of its offices at least 25% of the time, though some roles may require more office time. The position requires a bachelor’s degree or equivalent combination of education, training, and/or experience in a relevant field. Experience requirements will correlate with internal job-level requirements.

Requirements

Track record building and operating production systems in a hands-on engineering role with a DevOps flavor.
Production-quality Python and SQL skills; pipeline code is primarily Python, and the presentation layer uses BigQuery SQL, including table-valued functions and views.
Deep experience operating at least one major cloud provider: Amazon Web Services, Google Cloud, or Microsoft Azure.
Experience with observability tooling, including Prometheus, PromQL, and Grafana; ability to write recording rules and build monitoring relied upon by engineering teams.
Ability to gather requirements independently and work across organizational boundaries in an ambiguous environment with limited direction.
Bachelor’s degree or an equivalent combination of education, training, and/or experience.
A field of study relevant to the role, as demonstrated through coursework, training, or professional experience.

Nice To Haves

Experience with capacity planning, resource management, or cost attribution systems at a hyperscaler or in a large-scale machine learning environment; product engineering and developer experience are also relevant.
Experience with scheduling and packing efficiency, or profiling-driven optimization of large distributed workloads.
Multi-cloud data ingestion experience, especially normalizing billing exports, reservation APIs, on-demand capacity reservations, commitments, and provider telemetry across different billing arrangements.
Experience with total cost of ownership and forecasting, including distinguishing whether infrastructure growth is causal or correlated with business drivers.
Familiarity with accelerator infrastructure, including GPU metrics such as DCGM, TPU utilization, Trainium power and utilization metrics, or machine learning training and inference systems at the hardware level.
Experience building internal data products with self-service access, schema contracts, API serving, documentation, and discoverability.
Experience with storage efficiency, retention, and lifecycle programs at exabyte scale.

Responsibilities

Build planning and allocation tools used by leadership, teams, and schedulers, including cross-region and cross-provider placement, guardrails, queueing, and occupancy KPIs.
Drive efficiency programs involving stranding and rightsizing, recovery of unused capacity, and job-level utilization across training, inference, and evaluation workloads.
Establish per-configuration efficiency baselines and work with teams that own systems to improve utilization.
Reconcile billing across more than ten providers against telemetry and internal systems, attribute spend to workloads, and turn demand signals and research roadmaps into a defensible compute plan and supply pipeline.
Build and own data pipelines that ingest occupancy, utilization, and cost data from a diversifying fleet into BigQuery, with responsibility for completeness, latency SLOs, and gap detection.
Operate Kubernetes-native systems at scale, including collection agents, workload labeling, and taint, reservation, and scheduling behavior that determines usable capacity.
Gather requirements, define schema contracts, and design data products for consumers ranging from research engineers to finance leadership.
Provide on-call ownership and SLOs for load-bearing capacity data and tooling surfaces.
Build observability tooling and performance instrumentation that provide visibility into fleet health and workload hardware utilization.