Qualcomm
Senior AI Performance Architect
$127K–$190K
San Diego, California +4 · On-site · Senior · Full-time · Posted Oct 28, 2026
- Location
- San Diego, California
Austin, Texas
Folsom, California
Raleigh, North Carolina
Santa Clara, California - Experience
- 3+ years in hardware engineering defining architecture of GPUs or accelerators used for AI training. The posting also lists minimum qualification alternatives: bachelor's degree plus 2+ years of relevant work experience; master's degree plus 1+ year; or a PhD.
- Education
- Master's degree in Computer Science, Engineering, Information Systems, or related field. Minimum qualifications also list bachelor's or master's degree in these fields, or a PhD.
About the job
Architects and evaluates AI accelerator hardware and systems to improve performance, power efficiency, and scalability for AI model training.
What you'll do
- Determine how ML trends and customer requirements affect hardware and software architecture; analyze existing accelerator and GPU architectures.
- Architect accelerator, memory, and scale-out/scale-up enhancements, including hardware/software co-design for AI training workloads.
- Build performance models and pre-silicon predictions; assess performance, area, and power trade-offs.
What they ask for
- At least 3 years of hardware engineering experience defining GPU or AI-training accelerator architecture.
- In-depth knowledge of nVidia/AMD GPU capabilities and architectures.
- Knowledge of LLM architectures and their hardware requirements.
- Master's degree in Computer Science, Engineering, Information Systems, or a related field.
Nice to have
- Knowledge of computer architecture, digital circuits, hardware simulators, communication protocols, NoC designs, and memory technologies.
- Experience modeling hardware and workloads for performance and power estimates; high-level hardware modeling experience.
Pay
$126,700.00 - $190,100.00 USD base salary
Tools
GEMM, GEMV, NoC, SoC, C++, Python, DeepSpeed, FSDP
AI focus
AI accelerator architecture for training, including LLM hardware requirements, AI training systems, workload performance modeling, and hardware/software co-design.
Model-written summary · the original posting is authoritative