Principal Product Manager, Hardware
$338,000–$438,000 year
HybridSan Francisco, California, United States or Bellevue, Washington, United States
Job Summary
Own the GPU fleet lifecycle as a product, defining platform introductions from NVIDIA B200/H200 classes through next-generation systems and configuring node/cluster options for 64 to 1,024+ GPUs. Lead end-to-end new platform launches, aligning roadmaps with NVIDIA and ODM partners while translating demand signals into fleet investment recommendations defended before executives. Specify InfiniBand fabric configurations to ensure NCCL and MPI performance, partnering with engineering teams to convert silicon roadmaps into reliable SKUs. Win adoption across engineers, designers, and customers through data-driven writing and direct conversation, then ship, measure, and iterate based on benchmark results. Requires 7+ years in technical infrastructure product management, with 4 days in Bellevue or San Francisco.
Required Qualifications
- 7+ years of product management experience on technical infrastructure, hardware, systems, or platform products
- 10+ years of product management experience
- Ownership of a multi-team or multi-product roadmap
- Ability to turn data and customer signal into a clear decision about what to build next
- Track record of getting engineers, executives, and external partners to adopt a plan because they made them want to, not because you outranked them
- Experience shipping products that required coordinating hardware, software, and operations teams against hard external deadlines
- Fluency in technical conversations with hardware and systems engineers
- Comfort discussing GPU architectures, interconnects, memory, and data center constraints
- Experience making hardware, capacity, or fleet investment recommendations that committed significant capital under uncertainty
- Ability to write documents that lead with the conclusion and the evidence, not the background
- Ability to define iterative plans that move an organization from the current state towards the desired outcome
- Presence in the Bellevue or San Francisco office location 4 days per week
- Availability for Tuesday as the designated work from home day
Desired Qualifications
- Deep background in GPU hardware, systems, or the semiconductor ecosystem
- Direct experience in GPU or accelerator hardware, or systems and HPC (high-performance computing) product management
- Experience working inside the semiconductor or original equipment manufacturer (OEM)/ODM ecosystem
- Experience working directly with NVIDIA or another silicon vendor on roadmap alignment
- Experience owning cloud infrastructure capacity planning or fleet economics
- Hands-on familiarity with distributed training infrastructure, including InfiniBand fabrics and NCCL
- Experience running benchmark programs and using the results to drive buy or configure decisions
Hiring someone like this?
Get your role in front of qualified candidates on Sorce.