Senior GPU Performance Software Engineer

Intel Corporation•Santa Clara, CA
•$195,200 - $275,580•Hybrid

About The Position

The Software and AI organization is seeking a software engineer to develop and low-level optimize oneDNN, a complex, cross-platform, open-source performance library that supports deep learning applications. This is a low-level software engineering and hardware-acceleration role focused on optimized math primitives, parallel algorithms, and GPU kernels that power AI frameworks such as OpenVINO, TensorFlow, PyTorch, and ONNX Runtime on Intel hardware. It does not involve building, training, or tuning machine-learning models. The role includes developing high-performance GEMM, convolution, and attention kernels; designing GPU kernel generation infrastructure; optimizing memory traffic, fusion, mixed-precision and quantized execution; and profiling GPU primitives and runtime paths. The engineer will also help co-design GPU primitives for next-generation Intel GPUs, collaborate with hardware and compiler teams, and improve validation, benchmarking, and continuous-integration infrastructure. The position is based in the United States, with primary locations in Hillsboro, Oregon, and an additional location in Santa Clara, California. It is eligible for Intel's hybrid work model, splitting time between the assigned Intel site and off-site work. The advertised annual salary range is $195,200–$275,580 USD, with pay determined by location and job-related factors.

Requirements

  • BSc, MSc, or PhD in Computer Science, Computer Engineering, Mathematics, Physics, or a highly technical related field.
  • At least 5 years of professional software development experience with expert-level modern C++.
  • At least 2 years of hands-on GPU programming and kernel optimization experience using SYCL/DPC++, OpenCL, CUDA, or HIP; alternatively, at least 5 years of similar low-level performance optimization experience on CPUs.
  • Strong foundations in computer architecture, cache hierarchies, memory subsystems, and parallel programming paradigms such as multithreading and SIMD/vectorization.
  • Demonstrated ownership, initiative, and ability to resolve complex, ambiguous technical problems.
  • Ability to collaborate across hardware, compiler, and framework teams.
  • Performance-driven curiosity and interest in low-level systems optimization.

Nice To Haves

  • Experience developing high-performance math libraries, such as GEMM, convolution, reduction, or FFT kernels.
  • Hands-on experience with GPU assembly-level tuning or compiler optimization.
  • Familiarity with parallel programming APIs such as OpenMP or oneTBB.
  • Basic understanding of deep-learning primitives, such as forward and backward passes, and how library code is used by upstream frameworks.

Responsibilities

  • Develop high-performance GEMM, convolution, and attention kernels for AI workloads.
  • Design scalable JIT and code-generation infrastructure for GPU kernel generation.
  • Implement fusion and memory-traffic optimizations to maximize hardware utilization.
  • Optimize mixed-precision and quantized execution paths, including BF16, FP16, INT8, FP8, and FP4.
  • Build analytical and empirical performance models for kernel dispatch and tuning.
  • Profile and eliminate performance bottlenecks across oneDNN GPU primitives and runtime paths.
  • Co-design GPU primitives and kernel architectures for next-generation Intel GPUs.
  • Partner with hardware and compiler teams to shape future accelerator capabilities and software stacks.
  • Improve validation, benchmarking, and CI infrastructure for performance-critical GPU workloads.

Benefits

  • Competitive compensation and pay.
  • Stock programs and stock bonuses.
  • Quarterly bonuses.
  • Healthcare and health benefit programs.
  • Retirement benefits.
  • Vacation benefits.
  • Highly flexible hybrid/remote working options.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service