Senior Kernel Developer

Luxoft; United States; Remote; USD 80,000 - 100,000; Full time

Job summary

An innovative company is seeking an AI software development engineer to join a talented team focused on optimizing Machine Learning GPU kernels. In this exciting role, you will develop ML kernels in the Triton kernel language, directly impacting the performance of AMD GPUs and contributing to solutions for generative AI. Your expertise in C/C++ and CUDA or HIP programming will be crucial as you collaborate with the GPU architecture team to enhance future generations of technology. If you're passionate about cutting-edge technology and eager to make a significant impact, this opportunity is perfect for you.

Qualifications

• Proficiency in C/C++ and CUDA or HIP / ROCm programming is essential.
• Strong problem-solving skills and collaborative work experience required.

Responsibilities

• Develop ML kernels for matrix multiplication and Flash Attention.
• Benchmark and optimize kernels to enhance performance.

Skills

C/C++

CUDA

HIP / ROCm

OpenCL

Parallel programming models

Problem-solving skills

Job description

Project description

Luxoft is looking for an AI software development engineer to develop ML kernels in the Triton kernel language. We are looking for an engineer who is passionate about optimizing Machine Learning GPU kernels and improving the performance of key applications and benchmarks. What you do directly impacts the performance of AMD GPUs and enables us to become a competitive solution for generative AI. Become a part of our high-impact and incredibly talented Triton kernels team.

Responsibilities

• Develop ML kernels for matrix multiplication, Flash Attention, and other ML operators.
• Benchmark, perform competitive analysis, and optimize kernels to improve performance.
• Collaborate with the GPU architecture team to improve future generations.
• Apply knowledge of software engineering best practices.

SKILLS

Must have

• Proficiency with C/C++.
• Proficiency in CUDA or HIP / ROCm or OpenCL programming.
• Solid understanding of parallel programming models and optimization techniques.
• Strong problem-solving skills and the ability to work in a collaborative environment.

Nice to have

• Familiarity with models like LLama, Mixtral, and Gemma is a plus.
• Knowledge of MLIR, LLVM, GPU assembly, and GPU architecture is a plus.
• Familiarity with PyTorch or JAX.