PhD ML Systems Intern — Accelerate AI on Ryzen NPU

Advanced Micro Devices, Inc.; Santa Clara (CA); Hybrid; USD 42,000 - 54,000; Full time

Job summary

Advanced Micro Devices, Inc. is seeking a PhD ML System Engineering intern/co-op to accelerate LLM/VLM workloads on Ryzen AI NPU. You’ll own a project, study time and memory usage, and propose optimizations. This role is based in San Jose, CA or Santa Clara, CA with hybrid or onsite work for a 40-hour week.

You’ll learn transformer mapping to tiled dataflow accelerators, profiling for compute- vs memory-bound behavior, and you’ll present findings to the engineering organization at term end.

Qualifications

• Currently completing a PhD in CS, CE, or EE.
• Knowledge of LLM and VLM architectures, attention, KV cache.
• Experience with compilers: graph lowering, fusion, scheduling.
• Strong Python and C++ skills; able to profile code.
• Familiarity with AMD AI Engine (AIE) or IRON kernels.

Responsibilities

• Tackle a project to accelerate LLM/VLM on Ryzen AI NPU.
• Write and tune AIE kernels with IRON for major operators.
• Present findings to the engineering team and document learnings.

Skills

LLM/VLM architectures

Python

C++

Compiler concepts

Profiling performance

Education

PhD student in CS/CE/EE

Tools

MLIR

IRON

AIE

Job description

Advanced Micro Devices, Inc. is seeking a PhD ML System Engineering intern/co-op to accelerate LLM/VLM workloads on Ryzen AI NPU. You’ll own a project, study time and memory usage, and propose optimizations. This role is based in San Jose, CA or Santa Clara, CA with hybrid or onsite work for a 40-hour week.

You’ll learn transformer mapping to tiled dataflow accelerators, profiling for compute- vs memory-bound behavior, and you’ll present findings to the engineering organization at term end.