GPU AI Performance Engineer
Core
Design, optimize, and accelerate AI workloads on GPU-based platforms, focusing on LLMs, transformer workloads, MoE models, and computer vision pipelines.
Role type
Senior IC GPU AI Performance Engineer
Builds
Optimized AI inference runtimes and GPU kernels for Qualcomm platforms
Domain
Semiconductor / AI Systems Engineering
Deliverable
production ML models
Required skills
GPU architecture analysis, C/C++ programming, Python scripting, neural network layer optimization, compiler IR integration, low-level kernel optimization
Preferred skills
LLM inference engines (llama.cpp), GPU compiler pipelines (Triton, TTIR), quantization formats, MoE architectures, GPU driver development, OpenCL/Cuda development
Technologies
TensorFlow, PyTorch, Verilog, SystemVerilog, OpenCL, Cuda
Responsibilities
Analyze GPU microarchitecture for AI workload performance; Work across the AI stack from model graphs to GPU kernels; Collaborate with hardware/software/ML teams to identify bottlenecks; Analyze AI workload characteristics across GPU generations; Contribute to architectural trade-off studies and GPU roadmap decisions
Seniority
Senior, hands-on IC
