AI Performance Library Architect
Core
Design, development, and maintenance of new functionality in oneDNN, a cross-platform open-source library for neural network performance optimization.
Role type
Senior IC software development engineer (AI performance library)
Builds
oneDNN library, powering OpenVINO, Tensorflow, PyTorch, ONNX Runtime, and other AI frameworks
Domain
High-performance computing (HPC) and AI infrastructure
Deliverable
production ML models
Required skills
C and C++, software libraries design and architecture, linear algebra algorithms (BLAS, LAPACK, PyTorch), performance engineering, floating point arithmetic, Linux development, low-level optimizations (CUDA, x86 assembly, intrinsics, OpenCL)
Preferred skills
Machine learning and deep learning algorithms, HPC applications development, transcendental function implementations, non-IEEE low precision data types (bfloat16, fp8, fp4), AI assisted software development
Technologies
oneDNN, OpenVINO, Tensorflow, PyTorch, ONNX Runtime, CUDA, x86 assembly, OpenCL, BLAS, LAPACK
Responsibilities
Design and develop new functionality in oneDNN to enable performance critical portions of AI workloads; support software developers optimizing AI frameworks for Intel CPUs and GPUs; contribute to the cross-platform ecosystem of AI software developers
Seniority
Senior, hands-on IC