Member of Technical Staff, AMD GPU Performance Engineering
Core
Build and optimize AMD GPU backends, kernels, runtime paths, and benchmarking infrastructure for the vLLM AI inference engine to deliver frontier inference performance on AMD GPUs.
Role type
Senior IC GPU performance engineer (AMD ecosystem) (via careerplan.io/jobs/b96d4f8b-257b-4a82-8975-57e7fd885d87-member-of-technical-staff-amd-gpu-performance-engineering-at-inferact)
Builds
vLLM inference engine optimized for AMD accelerators
Domain
AI inference systems, GPU hardware, compilers, and kernel optimization
Deliverable
production ML models
Required skills
AMD GPU workload optimization, ROCm, HIP, Triton, CK, AITER, kernel performance tuning, memory behavior analysis, performance profiling, benchmarking, attention/GEMM/sampling/KV cache optimization
Preferred skills
vLLM, SGLang, TensorRT-LLM, compiler technologies (MLIR, LLVM), quantization methods (INT8, FP8, mixed precision), automated performance regression detection
Technologies
ROCm, HIP, Triton, CK, AITER, PyTorch, MLIR, LLVM
Responsibilities
Build and optimize AMD GPU backends and kernels; improve performance-critical paths (attention, GEMM, sampling, KV cache); develop benchmarking infrastructure; profile and measure hardware performance; optimize runtime paths for inference systems
Seniority
Senior, hands-on IC
