At Wafer, we are building AI systems that automatically optimize inference workloads across silicon. Turning better performance into a better business takes more than faster software. We need the right compute, on the right terms, serving the right demand. We're looking for someone who understands compute deeply and can connect the technical details to the economics. You should be able to go from GPU architecture to a supplier contract to a financial model, and explain what we should do. On-site in San Francisco, five days a week. Small team with massive surface area and ownership. You operate with complete autonomy of how to solve problems. You will work directly with engineering, finance and suppliers. Your work should end in better decisions about the compute we run, what it costs and how we scale.
Responsibilities
Work with engineering and finance to understand which models and hardware configurations make money, and what changes that; Build models connecting inference performance, utilization, pricing, compute costs and cash requirements; Evaluate GPU supply and supplier proposals, comparing price, availability, contract length, deposits, reliability and flexibility; Help source and negotiate capacity, then track whether delivery and performance match what was agreed; Manage provider relationships day to day, including onboarding new capacity, escalating incidents and holding providers to their commitments; Turn changing demand and hardware economics into concrete capacity decisions; Keep an accurate picture of the fleet: what is owned, where it is, what it costs, when contracts end and how well it is utilized; Investigate discrepancies across usage, invoices and contracts, and build systems that keep those numbers reliableRequirements
Understands compute beyond the spec sheet, including how GPU architecture, memory, networking and software affect the cost of serving a real workload; Is exceptional with numbers, can build a clear financial model, challenge its assumptions and explain the difference between profit and cash; Has roughly 1-5 years of relevant experience; Is credible with engineers and commercially sharp with suppliers, asking precise questions and following details through to a decision; Works without a spec, can take a half-defined problem, find the information that matters and come back with a concrete answerNice to have
A background in compute infrastructure, semiconductors, finance or investingBenefits
Fully covered medical, dental and vision insurance; Daily lunch and dinner; Unlimited PTO and parental leave; $1K/month post-tax housing stipend if living within 0.5 miles of the office; Relocation allowance and benefits; Covered Uber/Waymo from/to office; Visa sponsorship available; EquityHow to apply
Apply online: https://www.jobs.pl/en/oferta/102582364/aplikuj
JOBS.pl · https://www.jobs.pl/en/operations-compute-oferta-102582364




