Head of Compute

Другое

• We are building a new set of financial solutions for the companies that build, own, operate, and finance AI compute. These are complex solutions, and they only work if the people designing them understand the underlying hardware as well as the people who run it.
• We want one person who understands compute deeply enough to shape this business: someone who can look at a fleet, a facility, or a contract and tell us what it is really worth, what could go wrong, and how to build something around it that customers and capital partners can rely on.
• What you will own
• Compute economics. You own our view of what compute costs and what it is worth over time. That means real modeling rather than headline numbers: total cost of ownership across hardware generation, configuration, location, power, utilization, and contract structure. When someone asks whether a number holds up, you are the person who knows.
• Solution design. You work with our product, underwriting, and capital teams to design and structure financial solutions around compute assets and commitments. You translate how hardware is deployed, used, and retired into the terms and protections a financial solution needs.
• Diligence and risk. You evaluate what sits behind each opportunity: the hardware and its generation, the facility, power and cooling, utilization history, the contract terms, and the strength of the operator. You know which operators deliver on their commitments and which do not.
• Hardware lifecycle. You own our view of how compute assets age: useful life by generation, the effect of each new release, refresh cycles, and what equipment is worth at each stage of its life.
• Relationships. You are our credible technical voice with cloud providers, GPU specialists, datacenter operators, hardware vendors, and financing partners.
• What you have done
• Know GPU infrastructure at scale and can talk about specific systems by generation: A100, H100, H200, B200, B300, and what is different about each in deployment, performance, power, and useful life.
• Built or run large GPU fleets, or worked closely enough with them to know where the cost, the risk, and the failure modes live: networking, scheduling, utilization, and hardware failure.
• Worked on the commercial or financial side of compute: capacity contracts, cost and depreciation models, equipment financing or leasing, or investment diligence on AI infrastructure.
• Explained technical detail in financial terms to people who make decisions on it: investors, lenders, underwriters, or executives.
• Led teams and cross-functional programs. Twelve or more years of experience, with at least five leading teams or leading leaders.
• What makes you unusual
• Most people in this market are one of two things: an engineer who has never had to justify a dollar, or a financier who has never run a node. We want the person who is credible in both rooms. If you have ever looked at a deal and known within a minute whether the compute behind it was sound, and could then explain why to a room of investors, you are the person we are describing.
• Useful but not required: time at a hyperscaler, GPU cloud, or datacenter operator; experience in infrastructure investing, equipment finance, project finance, or insurance; time at a national lab or HPC center.
• How we will evaluate you
• We will give you a real compute opportunity to take apart: the hardware, the operator, and the contract. We will ask you what it is worth, what could go wrong, and how you would build a solution around it. We care far more about the quality of those answers than about pedigree.