Staff Engineer, Inference on Multi-Hardware

RadixArk; Palo Alto (CA); Presencial; USD 200 000 - 400 000; Tempo integral

Resumo da oferta

RadixArk is hiring a Member of Technical Staff - Inference-Multi-Hardware to push frontier AI performance across NVIDIA/AMD GPUs, TPUs, CPUs, and emerging accelerators. You will bring up SGLang, Miles, and the infrastructure, optimizing kernels and runtimes for multiple hardware platforms.

You will design abstractions to keep a single codebase fast across vendors, port kernels, and drive performance end-to-end with vendors and silicon partners during pre-release stages.

Qualificações

• 4+ years of experience in systems, performance, or ML infrastructure engineering
• Deep expertise in at least one accelerator programming model (CUDA, ROCm/HIP, Pallas/XLA, Triton, or a vendor SDK)
• Strong understanding of accelerator architecture: memory hierarchy, bandwidth limits, occupancy
• Experience writing or optimizing high-performance kernels for ML workloads
• Experience with distributed execution and communication libraries (NCCL, RCCL, MPI, or equivalents)
• Proficiency in C++ and Python
• Strong debugging and profiling skills at the system level
• Track record of performance work that shipped into production

Responsabilidades

• Bring up RadixArk's inference and training systems on new accelerator platforms and drive them to competitive performance
• Design hardware abstractions that let a single codebase stay fast across vendors without forking
• Port and optimize kernels across programming models and memory architectures
• Build cross-platform benchmarking, profiling, and regression detection so performance claims hold up on every target
• Debug numerical divergence and correctness gaps between platforms
• Work with vendor engineering teams on pre-release hardware, compiler and driver issues
• Partner with kernel, runtime, distributed systems, and product engineers to land performance wins end to end
• Serve as the internal source of truth on what each platform is actually good at
• Contribute hardware-specific optimizations, benchmarks, and portability work back to open-source SGLang and Miles

Conhecimentos

C++

Python

Performance engineering

Kernel optimization

Distributed systems

Profiling tools

Accelerator programming

ML infrastructure

Debugging

Production-grade performance

Ferramentas

NCCL

RCCL

MPI

XLA

MLIR

TVM

Triton

SGLang

Miles

CUDA

ROCm/HIP

oneDNN

AVX-512

AMX

Descrição da oferta de emprego

RadixArk is hiring a Member of Technical Staff - Inference-Multi-Hardware to push frontier AI performance across NVIDIA/AMD GPUs, TPUs, CPUs, and emerging accelerators. You will bring up SGLang, Miles, and the infrastructure, optimizing kernels and runtimes for multiple hardware platforms.

You will design abstractions to keep a single codebase fast across vendors, port kernels, and drive performance end-to-end with vendors and silicon partners during pre-release stages.