Hardware Systems Architect (AI)

TL;DR

Hardware Systems Architect (AI): Designing and optimizing the system-level architecture for large-scale AI compute fleets with an accent on interconnects, power, cooling, and rack-level integration. Focus on developing high-level specifications, evaluating hardware trade-offs, and scaling infrastructure for frontier AI models.

Location: Hybrid (San Francisco, New York City, or Seattle). Staff are expected in the office at least 25% of the time.

Salary: $320,000 – $485,000 USD

Company

Anthropic is a public benefit corporation dedicated to creating reliable, interpretable, and steerable AI systems that are safe and beneficial for society.

What you will do

• Own system path-finding and architecture for compute deployments from initial concept and requirements through spec, design review, and deployment.
• Author high-level requirements and specifications for boards, interfaces, and racks that internal teams and partners build against.
• Drive interconnect and networking architecture, including scale-up and scale-out fabrics, NICs, switches, and topology.
• Direct board-, chassis-, and system-level architecture, focusing on component placement, PCB architecture, and thermal budgeting.
• Develop performance simulations and hardware models to evaluate and select optimal hardware architectures.
• Define validation, bring-up, and qualification strategies for new platforms, including telemetry and reliability targets.

Requirements

• Deep, hands-on expertise in core hardware-systems domains (e.g., interconnect, power, thermal) for large-scale compute or networking.
• Experience owning hardware system architecture at scale (machine, rack, row, and cluster) and authoring technical specifications.
• Track record of taking hardware from the architecture phase through bring-up and high-volume production deployment.
• Experience managing external hardware vendors, reviewing designs, and ensuring adherence to specifications.
• Ability to reason about trade-offs across adjacent hardware domains rather than within a single specialization.
• Visa sponsorship is available, though it is not guaranteed for every role.

Nice to have

• 8+ years in hardware systems architecture for hyperscale, HPC, AI/ML, or high-end networking platforms.
• Experience with AI accelerator systems (GPU, TPU, or custom ASICs) and their associated fabrics.
• Familiarity with chip-package-system interface and co-design.
• Specialized depth in high-speed SerDes, PCIe/CXL, Ethernet/InfiniBand, or direct liquid cooling.
• Experience establishing new engineering functions or programs.

Culture & Benefits

• Collaborative "big science" environment focusing on high-impact research efforts.
• Competitive compensation with optional equity donation matching.
• Generous vacation and parental leave policies.
• Flexible working hours and access to high-quality collaborative office spaces.
• Commitment to diversity and representation in the development of AI systems.