Data Center Engineer, Reliability & Infrastructure Management – Compute Supply

About Anthropic Anthropic’s mission is to create reliable, interpretable, and steerable AI systems. We want AI to be safe and beneficial for our users and for society as a whole. Our team is a quickly growing group of committed researchers, engineers, policy experts, and business leaders working together to build beneficial AI systems

About the role

As an engineer on this team, you'll sit at the intersection of IT and facilities. You'll build the power models, availability models, topology and load management requirements that operate more efficiently against our physical envelope. Your work will directly shape which sites we lease, what SLAs we sign, how far we oversubscribe, and where workloads are placed

What you'll do

• Topology and load management.
• Own the power and cooling topology of our fleet as a living dataset, validate it and its telemetry against how the buildings really behave, and define how load management behaves under failure across chips and capacity providers, from requirements through commissioning and incident support.
• Reliability modeling.
• Build and audit availability models of data center electrical and mechanical systems, define Cloud availability zones and failure domains, from a single site to the whole fleet.
• Power modeling and load forecasting.
• Build models of power draw from chip to rack to facility, by workload and hardware generation, and use them to set capacity planning targets, design envelope, and load forecasting.
• Partner diligence.
• Work with data center developers, operators, cloud providers and chip vendors to drive design improvements and hold them to SLA-grade performance, with technical diligence on their architectures, reliability studies, rack power specifications and operator interfaces