Staff DevTech: Accelerate AI Inference on GPUs

RadixArk; Palo Alto (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

RadixArk is seeking a Member of Technical Staff to accelerate LLM inference and training on modern GPUs. You will profile, optimize, and extend SGLang and Miles across production workloads, collaborating with partner teams to push the performance envelope and broaden model support.

You will work on kernel development, day-0 model support, and enablement for new silicon, while contributing to the roadmap and ecosystem integrations that power scalable AI inference and training.

Qualifications

• 4+ years of experience in GPU systems, LLM infrastructure, or performance engineering.
• Strong profiling and debugging skills across the stack.
• Hands-on GPU programming in CUDA, ROCm, or Triton.
• Strong Python and C++ or CUDA programming.
• Ability to translate ambiguous asks into technical plans and recipes.
• Excellent communication with engineering teams.

Responsibilities

• Profile and optimize GPU performance for real production workloads on current and next-generation hardware.
• Go deep in one or two focus areas: inference performance, kernels, model/hardware enablement, speculative decoding, training systems.
• Partner directly with the ecosystem to turn ambiguous problems into concrete wins and reproducible cookbooks.
• Enhance SGLang and Miles by feeding user-driven improvements back into core systems.

Skills

GPU systems

LLM infrastructure

Profiling & debugging

Python

C++/CUDA

Cross-platform

Communication with engineers

Tools

CUDA

ROCm

Triton

Job description

RadixArk is seeking a Member of Technical Staff to accelerate LLM inference and training on modern GPUs. You will profile, optimize, and extend SGLang and Miles across production workloads, collaborating with partner teams to push the performance envelope and broaden model support.

You will work on kernel development, day-0 model support, and enablement for new silicon, while contributing to the roadmap and ecosystem integrations that power scalable AI inference and training.