Principal Software Engineer - Performance Tooling
Core
Debug, profile, and optimize performance for large-scale AI model training and inference workloads across the software stack (abstractions, compilers, runtimes, libraries, APIs) on CPUs and GPUs.
Role type
Principal Software Engineer (AI Performance Tooling)
Builds
Scalable, production-ready AI performance improvements for training and inference
Domain
Artificial Intelligence / High Performance Computing
Deliverable
production ML models
Required skills
C++, Python, GPU architecture, CPU architecture, DNN/LLM inference, DL frameworks (PyTorch, Tensorflow, ONNX Runtime), CUDA, ROCm, Triton, GPU profiling tools, high performance application development, performance debugging, performance optimization
Preferred skills
Cross-team collaboration, independent project leadership
Technologies
PyTorch, Tensorflow, ONNX Runtime, CUDA, ROCm, Triton
Responsibilities
Debug and profile performance for training/inference workloads on CPUs/GPUs, Monitor performance regressions, Drive continuous improvements to reduce time-to-deploy and hardware footprint, Collaborate with researchers and engineers to deliver scalable AI performance improvements, Optimize state of the art LLMs and HPC applications, Utilize GPU profiling tools for end-to-end performance analysis
Seniority
Principal, hands-on IC with strategy & mentorship
