SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Anthropic is building reliable, interpretable, and steerable AI systems. This role sits within Capacity Engineering's Planning pillar on the Demand Planning team, owning the critical bridge between infrastructure demand forecasts and physical capacity delivery across Anthropic's massive, multi-region accelerator fleet (spanning multiple cloud providers, neoclouds, and on-premises sites).
You will own the end-to-end "tranche" lifecycle—converting demand forecasts into concrete per-tranche requirements (accelerator type, interconnect, region, supporting resources, delivery date), representing those requirements in sourcing negotiations and data center build reviews, and then tracking every tranche from contract signature through reserved, ingested, in-cluster, healthy, and occupied states. You'll define the canonical state machine for capacity bring-up, qualify tranches for deliverability before signature, and close the delivery loop by publishing forecast-versus-actual variance and feeding it back into planning.
Key responsibilities include: translating demand forecasts into per-tranche infrastructure requirements; qualifying tranches for deliverability across shape, region, and timing; owning the integrated schedule and system of record for all in-flight capacity; running a portfolio of parallel bring-ups (new cloud regions, on-prem sites, neocloud blocks) with one integrated schedule; driving readiness automation across all capacity systems; and publishing executive-level reporting on time-to-occupied, paid-idle dollars, status, tradeoffs, and risk.
You'll partner daily with research engineering, pretraining, inference, compute supply, finance, and external vendors. The role requires significant experience delivering large-scale infrastructure at multi-region scale or with ≥10k accelerators, technical depth from cluster orchestration and node health up through telemetry and planning systems, SQL and Python proficiency, and a technical degree or equivalent engineering track record. Preferred experience includes reserved-capacity onboarding, demand-planning exposure, data center delivery, accelerator health and burn-in, systems of record for infrastructure assets, and onboarding new hardware generations.
