PhD AI Systems & GPU Performance Intern (Co‑op)

AMD; San Jose (CA); Hybrid; USD 60,000 - 80,000; Full time

Benefits offered by this job

AMD Benefits

Job summary

AMD seeks a highly motivated PhD candidate for an AI Systems & GPU Performance Engineering Intern role in the San Jose area. You will optimize state-of-the-art AI models, training workflows, and inference apps on AMD Instinct GPUs using ROCm, PyTorch, JAX, and related tools.

You will profile, benchmark, and develop reproducible benchmarking frameworks, aiming to improve efficiency, scalability, throughput, and latency across systems.

Qualifications

• Pursuing a PhD with expected graduation after internship.
• Strong programming in Python and C/C++, with accelerator frameworks exposure.
• Experience with DL frameworks (PyTorch/JAX/TF) and large-scale model training concepts.
• Linux scripting, debugging, and performance profiling.
• Familiar with GPU architecture concepts and profiling approaches.

Responsibilities

• Profile, benchmark, and optimize AI training and inference workloads on AMD Instinct GPUs.
• Develop reproducible benchmarking frameworks and automation for performance validation.
• Analyze and optimize end-to-end AI workflows, including LLMs and generative AI, using quantization and mixed precision.
• Collaborate with engineers/researchers to evaluate GPU performance across environments.
• Contribute data-driven recommendations to improve efficiency, throughput and latency.

Skills

Python

C/C++

CUDA

HIP

Triton

PyTorch

JAX

TensorFlow

Linux

Profiling

Performance analysis

GPU architecture

Education

PhD in CS/CE/Applied Math

Tools

rocProfiler

ROCm System Profiler

Omniperf

Nsight

Job description

AMD seeks a highly motivated PhD candidate for an AI Systems & GPU Performance Engineering Intern role in the San Jose area. You will optimize state-of-the-art AI models, training workflows, and inference apps on AMD Instinct GPUs using ROCm, PyTorch, JAX, and related tools.

You will profile, benchmark, and develop reproducible benchmarking frameworks, aiming to improve efficiency, scalability, throughput, and latency across systems.