Software Engineer – GPU Kernel

What you'll do

• Build and optimize inference kernels, including GEMM, attention, routing, reduced-precision, and quantized kernels.
• Develop GPU code in CUDA and C++, benchmark performance across NVIDIA and AMD hardware, and tune performance-critical components.
• Accelerate multi-modal model pipelines and integrate next-generation GPU features.

What they ask for

• 3+ years of experience in GPU programming, HPC, or performance-critical systems.
• Bachelor’s or Master’s degree in Computer Science, Computer Engineering, Electrical Engineering, or a related field.
• Strong proficiency in CUDA for NVIDIA GPUs or ROCm/HIP for AMD GPUs, and proficiency in C++.
• Understanding of GPU architecture, profiling and performance tuning, numerical precision trade-offs, and quantization techniques.

Nice to have

• Kernel-level optimization of transformer, multi-modal, or Mixture-of-Experts (MoE) architectures; familiarity with CUTLASS or Triton.
• Inter-GPU communication experience, relevant open-source contributions, or research/conference presentations in GPU optimization, HPC, or numerical computing.

Tools

CUDA, C++, ROCm/HIP, FP8/FP4, CUTLASS, Triton

AI focus

Develop and optimize GPU kernels for AI inference, including quantized and reduced-precision computation, multi-modal pipelines, and model architectures such as transformers and MoE.