RadixArk
Member of Technical Staff — Inference-Kernel, Compiler & Communication
$200K–$400K
Palo Alto +1 · On-site · Staff · Full-time · Posted Feb 17, 2026
- Location
- Palo Alto
San Francisco - Experience
- 5+ years in systems, compiler, or performance engineering.
About the job
Develop and optimize low-level kernels, compilers, runtimes, and communication systems to improve performance for distributed AI training and inference.
What you'll do
- Design and implement high-performance kernels and optimize compiler and runtime stacks for ML systems.
- Improve communication efficiency, latency, and throughput across large GPU clusters.
- Profile system bottlenecks and build performance-analysis tools in collaboration with training and inference teams.
What they ask for
- Strong expertise in CUDA or accelerator programming, GPU architecture, and memory hierarchy.
- Experience writing or optimizing high-performance kernels and working with compilers, runtimes, or code generation.
- Experience with distributed communication libraries and networking or interconnect technologies.
- Proficiency in C++ and Python, with system-level debugging and profiling skills.
Nice to have
- Experience with Triton, TVM, XLA, or MLIR.
- Experience optimizing collective communication at scale or scaling workloads to 1000+ GPUs.
Pay
$200,000 - $400,000 USD annual base salary
Tools
CUDA, C++, Python, NCCL, MPI, RCCL, Triton, TVM
AI focus
Performance engineering for AI training and inference, including kernel optimization and distributed workloads.
Model-written summary · the original posting is authoritative