Fellow, AI Performance Software Architect
Core
Architect and optimize AI software solutions for next-generation GPU computational accelerators (Instinct GPUs) to enable deep learning models, libraries, and applications in on-prem and cloud environments.
Role type
Senior AI Performance Software Architect (IC)
Builds
Custom AI software solutions, optimized inference/training frameworks, and system-level performance tools for data centers and supercomputers.
Domain
AI / High-Performance Computing / GPU Architecture
Deliverable
production ML models
Required skills
C++, Python, Deep Learning frameworks (inference/fine-tuning/training), GPU performance optimization, system-level architecture, root-cause analysis
Preferred skills
Profiling tools (Torchprofiler, RocM profiler, Vtune, Nsight), Parallel programming (NCCL/RCCL, OpenMP, MPI), Containerization (Singularity, Docker, Kubernetes), Open-source contribution
Technologies
Instinct GPUs, PyTorch, TensorFlow, ROCm, NCCL, MPI, Docker, Kubernetes, Vtune, Nsight
Responsibilities
Develop and optimize software for AI workloads on Instinct GPUs; analyze hardware bottlenecks to maximize performance; collaborate on open-source projects; implement parallel methods on GPU accelerators.
Seniority
Senior, hands-on IC
