PMTS Software Engineer - GPU Kernels for ML Acceleration

AMD; Bellevue (WA); On-site; USD 120,000 - 180,000; Full time

Job summary

AMD is seeking a PMTS Software Development Engineer to develop and optimize low-level GPU kernels for accelerating inference and training of large ML models. You will work on distributed model training across multiple GPUs and nodes, profiling performance and optimizing resource utilization to maximize efficiency.

Collaborate with ML researchers and infrastructure teams to productionize optimized kernels and contribute to documentation and best practices.

Qualifications

• Five years of experience in the job offered or closely related engineering role.

Responsibilities

• Develop and optimize low-level GPU kernels to accelerate inference and training of large ML models.
• Profile and optimize system performance across GPUs and nodes to maximize throughput.
• Collaborate with ML researchers and software engineers to integrate optimized kernels into production systems.
• Create detailed documentation of optimizations and guidelines for maintainable code.

Skills

Low-level GPU kernels

Distributed ML training

Performance profiling

Verilog RTL coding

UVM

VLSI Design

Python scripting

TCL/TK

FPGA

Education

Master’s degree in CS/CE/EE/SE

Bachelor’s degree + 7 years experience

Tools

CUDA

Job description

AMD is seeking a PMTS Software Development Engineer to develop and optimize low-level GPU kernels for accelerating inference and training of large ML models. You will work on distributed model training across multiple GPUs and nodes, profiling performance and optimizing resource utilization to maximize efficiency.

Collaborate with ML researchers and infrastructure teams to productionize optimized kernels and contribute to documentation and best practices.