Product Manager (AI Hardware)
Core
Design and execute performance benchmarks for AI accelerators (GPUs, TPUs, custom silicon) to measure throughput, latency, and price-performance, building evaluation methodologies and cost models for the industry.
Role type
Product Manager (AI Hardware)
Builds
Hardware benchmarking stack (AA-AgentPerf), performance analysis reports, and cost/throughput models for AI accelerators.
Domain
AI Infrastructure / Semiconductor Hardware
Deliverable
production ML models
Required skills
Python, data analysis, AI accelerator architecture, inference software stack (CUDA, TensorRT, vLLM), inference economics, critical thinking
Preferred skills
Experience with frontier AI models, relationships with chipmakers
Technologies
CUDA, TensorRT, vLLM, Python
Responsibilities
Design and run performance benchmarks across accelerators, develop evaluation frameworks, analyze inference economics, partner with chipmakers on benchmarking, produce strategic analysis reports.
Seniority
Mid-Senior, hands-on IC