Member of Technical Staff, Developer Technology (DevTech) at RadixArk

Department: Member of Technical Staff

Compensation

$200,000 – $400,000/yr

Location: Palo Alto, California
Type: Full-time
Level: senior
Posted: Jul 23, 2026

Description

RadixArk is hiring a Member of Technical Staff, Developer Technology (DevTech) to accelerate LLM inference and training on modern GPU hardware. The role focuses on profiling and optimizing GPU performance, building custom CUDA/ROCm/Triton kernels, enabling new models and hardware, supporting speculative decoding and training systems, and translating expert user problems into reproducible technical guidance and open-source improvements. Candidates need at least four years of experience in GPU systems, LLM infrastructure, or performance engineering, along with strong profiling, debugging, Python, C++, and GPU-programming skills.

Required Skills

GPU systems
LLM infrastructure
Performance engineering
CUDA
ROCm
Triton
Python
C++
Profiling
Debugging
Low-precision quantization
Speculative decoding
Distributed training

Additional Information

Required Previous Experiences: At least 4 years of experience in GPU systems, LLM infrastructure, or performance engineering.
Hands-on GPU programming experience with CUDA, ROCm, or Triton and willingness to work across platforms.
Preferred Qualifications: Deep familiarity with LLM inference internals, including distributed serving, parallelism, routing, KV-cache management, and scheduling.
Experience with low-precision quantization and inference or training, including FP8, INT8/INT4, NVFP4, or MXFP4.
Experience writing and optimizing custom GPU kernels.
Practical familiarity with speculative decoding methods such as Eagle, DFlash, or DSpark.
Working knowledge of large-scale distributed training, including pre-training, SFT, RL post-training, elasticity, and long-context workloads.
Experience optimizing across both NVIDIA and AMD platforms.
Hands-on experience with SGLang, Miles, vLLM, TensorRT-LLM, Megatron, or comparable frameworks, including contributions to open-source AI/ML projects.