✦ Luna Orbit — Software Engineering

Software Development Engineer - C++, GPU Math Libraries

at Advanced Micro Devices

📍 Austin, Texas, United States Hybrid Posted March 19, 2026
Apply Now → ← Browse All Jobs

Position Details

Type Full-Time
Experience mid
Exp. Years Entry to mid-level
Education Bachelor's orMaster'sdegree in Applied Mathematics, Computer Science, Computer Engineering, Electrical Engineering, or equivalent
Category Software Engineering

About this role

This role involves developing high-performance GPU kernels for machine learning libraries, focusing on GEMM operations and fusions, within AMD's ROCm ecosystem.

Key Responsibilities

  • Design and implement GPU kernel generator
  • Develop build and testing systems
  • Collaborate with teams on ML primitives
  • Optimize GPU code
  • Document best practices

Technical Overview

Technical environment includes C++, GPU programming, ROCm platform, high-performance computing, and open-source collaboration.

Ideal Candidate

The ideal candidate is a mid-level software engineer with strong expertise in C++, GPU programming, and high-performance computing. They should have experience developing and optimizing GPU kernels, particularly for machine learning libraries, and be comfortable collaborating across teams.

Must-Have Skills

C++GPU code optimizationGPU kernel developmentGEMMsoftware engineering methodologies

Nice-to-Have Skills

assembly programmingvectorizationmachine learning primitivesopen-source software development

Tools & Platforms

ROCmGCCbuild systemstesting systems

Required Skills

C++GPUGEMMkernel generatorsoftware developmentprofilingdebuggingassembly programmingvectorizationopen-source

Hard Skills

C++C++GPUGPU kernel generatorGEMMGEMM fusionsmodern C++software developmenttestingprofilingdebuggingversion controlassembly programmingvectorizationApplied MathematicsComputer ScienceComputer EngineeringElectrical Engineering

Soft Skills

collaborationcontinuous learningcommunicationteamworkproblem-solving

Industry & Role

Industry Semiconductors & Hardware
Job Function Developing high-performance GPU kernels for machine learning libraries
Role Subtype Software Engineer
Tech Domains Active Directory, Microsoft 365, Azure, Linux, Windows Server, Kubernetes, Docker, Python, Java, SQL / PostgreSQL

Keywords for Your Resume

C++GPUGPU kernel generatorGEMMGEMM fusionssoftware developmentprofilingdebuggingversion controlassembly programmingvectorizationopen-sourcemachine learning primitiveshigh-performance computingROCmAMDAIdata centerscollaborationcontinuous learningGPU code optimization

Deal Breakers

Lack of C++ experience, No GPU programming background, No experience with high-performance kernel development, Bachelor's degree not in relevant field

Apply for this Position →

Get matched to jobs like this

Luna finds roles that fit your skills and career goals — no endless scrolling required.

Create a Free Profile