Software Engineer – GPU Kernel

6 months ago

Responsibilities

Design, implement, and optimize high-performance GPU kernels for AI inference, including GEMM, attention, and routing.
Develop and maintain GPU code in CUDA and C++, including low-level assembly when needed.
Implement reduced-precision and quantized FP8/FP4 kernels for low-latency and high-throughput inference.
Benchmark NVIDIA and AMD hardware and ensure cross-vendor performance parity.
Contribute to internal GPU libraries and tune performance-critical components.
Accelerate multimodal model pipelines and integrate next-generation GPU features.
Collaborate with the platform team to deploy GPU kernel work into production.

At least 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.
Deep understanding of GPU architecture, including warps, threads, memory hierarchy, synchronization, and latency-throughput trade-offs.
Proficiency in C++.
Experience with GPU profiling and performance tuning.
Strong numerical background and understanding of precision trade-offs and quantization techniques.
Preferred: experience optimizing transformer, multimodal, or Mixture-of-Experts architectures at the kernel level.
Preferred: familiarity with CUTLASS and Triton.
Preferred: inter-GPU communication programming experience, open-source GPU performance or ML acceleration contributions, and research or conference presentations on GPU optimization, HPC, or numerical computing.

51-200 employees