Member of Technical Staff - Kernel Engineer (AI)

Location: Santa Clara, California, United States

Salary: $170,000–$350,000 USD per year

Company

Builds and operates a high-performance AI inference platform that helps enterprises run large language models faster, more efficiently, and at scale.

What you will do

• Design, implement, and optimize GPU kernels for GEMM, attention, normalization, quantization, KV-cache, and Mixture-of-Experts workloads.
• Develop hardware-aware optimizations using CUDA, Triton, C++, PTX, CUTLASS, and related GPU programming frameworks.
• Profile production inference workloads and resolve bottlenecks involving computation, memory, synchronization, data movement, and kernel launch overhead.
• Improve end-to-end LLM inference latency, throughput, concurrency, memory utilization, and hardware efficiency.
• Integrate optimized operators into PyTorch, Triton, vLLM, SGLang, TensorRT-LLM, or equivalent runtimes.
• Build benchmarks, correctness tests, and regression tests while optimizing multi-GPU workloads and supporting NVIDIA, AMD, and emerging accelerators.

Requirements

• Strong experience in GPU programming, kernel development, high-performance computing, ML systems, or performance engineering.
• Proficiency in C++ and hands-on experience with CUDA, Triton, or another accelerator programming model.
• Strong understanding of GPU architecture, including memory hierarchy, warps, blocks, grids, shared memory, registers, Tensor Cores, synchronization, occupancy, and instruction-level parallelism.
• Experience optimizing performance-critical machine-learning operations such as attention, GEMM, quantization, KV cache, MoE, or Transformer operators.
• Familiarity with PyTorch and modern ML inference or training stacks, plus strong knowledge of numerical correctness and mixed-precision computation.
• Strong debugging, performance-analysis, communication, and collaboration skills.

Nice to have

• Experience with PTX/SASS, CUTLASS, CuTe, CUB, Thrust, FlashAttention, FlashInfer, or related low-level GPU libraries.
• Experience with vLLM, SGLang, TensorRT-LLM, ROCm/HIP, or AMD GPU architectures.
• Experience with multi-GPU or multi-node systems, NCCL, collective communication, distributed inference, or distributed training.
• Knowledge of torch.compile, XLA, MLIR, TVM, or other compiler and runtime systems.
• Open-source contributions or experience bringing new model architectures and accelerator platforms into production.

Culture & Benefits

• Hands-on individual-contributor role with significant technical ownership.
• Close collaboration with inference, distributed-systems, infrastructure, hardware-vendor, and engineering teams.
• Equity or stock options.
• Medical, dental, and vision coverage.
• Unlimited paid time off.
• H-1B and other work visa sponsorship available.