Principal Software Development Engineer
Core
Lead performance optimization of large-scale AI training/RL/inference workloads on AMD GPU platforms, driving innovations across the full software-hardware stack to maximize throughput and efficiency.
Role type
Principal GPU performance engineering lead (IC)
Builds
High-performance AI training and inference systems on AMD GPU platforms
Domain
Semiconductor / High-performance computing / Generative AI
Deliverable
production ML models
Required skills
GPU architecture expertise, distributed systems optimization, system-level bottleneck analysis, kernel and runtime optimization, cross-stack optimization (kernels, compilers, runtimes, frameworks), advanced profiling and benchmarking, Python, C++/CUDA/HIP
Preferred skills
Experience with Megatron-LM, Torchtitan, vLLM, Sglang, NCCL/RCCL, compiler stacks, graph-level optimization
Technologies
ROCm, Nsight, PCIe, Infinity Fabric, RDMA, NCCL, RCCL, Megatron-LM, Torchtitan, vLLM, Sglang
Responsibilities
Lead performance optimization of large-scale AI workloads; Identify and eliminate system bottlenecks across compute, memory, and communication; Drive cross-stack optimizations; Develop and apply advanced profiling methodologies; Collaborate with hardware and framework teams to influence architecture; Contribute to open-source efforts
Seniority
Principal, hands-on IC with technical leadership