Principal AI Hardware Architect
Core
Lead performance analysis, profiling, and analytical modeling across GPU and AI accelerator architectures to identify bottlenecks and drive perf/W and TCO optimization.
Role type
Principal AI Hardware Architect
Builds
AI training and inference workloads on production-scale systems
Domain
Semiconductor hardware design and AI systems
Deliverable
production ML models
Required skills
GPU and AI accelerator architecture analysis, analytical performance modeling, silicon measurement correlation, kernel-level optimization, Python programming, C/C++ programming, AI workload characterization, root-cause analysis, distributed training/inference framework knowledge, quantization and sparsity techniques
Preferred skills
Experience with PyTorch, vLLM, SGLang, Flash Attention, KV-cache management, communication-computation overlap strategies
Technologies
PyTorch, vLLM, SGLang, C/C++, Python
Responsibilities
Lead performance analysis, profiling, and benchmarking across GPU and AI accelerator architectures; Analyze end-to-end AI workloads and serving systems to understand performance and scalability drivers; Develop performance and system-level models to evaluate architectural features and innovations; Correlate silicon measurements with architectural models to guide future design decisions; Design and develop data analysis and performance modeling tools; Partner with architecture, compiler, and systems teams to influence product roadmaps; Present performance findings and architectural recommendations to senior technical leadership
Seniority
Principal, hands-on IC with strategic influence
