Senior GPU Performance Software Engineer (AI)

Location: Hybrid work model in the United States, with on-site work at an assigned Intel site in Hillsboro, Oregon or Santa Clara, California.

Annual salary: $195,200–$275,580 USD

Company

hirify.global's Software and AI organization contributes to oneDNN, an open-source performance library that powers deep learning applications on Intel hardware.

What you will do

• Develop high-performance GEMM, convolution, and attention kernels for AI workloads.
• Design JIT and code-generation infrastructure for GPU kernel generation.
• Implement fusion, memory-traffic, mixed-precision, and quantized execution optimizations.
• Build performance models, profile GPU primitives and runtime paths, and eliminate bottlenecks.
• Co-design GPU primitives and kernel architectures with hardware and compiler teams.
• Improve validation, benchmarking, and CI infrastructure for performance-critical workloads.

Requirements

• BSc, MSc, or PhD in Computer Science, Computer Engineering, Mathematics, Physics, or a related technical field.
• 5+ years of professional software development experience with expert-level modern C++.
• 2+ years of GPU programming and kernel optimization with SYCL/DPC++, OpenCL, CUDA, or HIP, or 5+ years of comparable low-level CPU performance optimization.
• Strong knowledge of computer architecture, cache hierarchies, memory subsystems, and parallel programming, including multithreading and SIMD/vectorization.
• Strong ownership, collaboration, and performance-focused problem-solving skills.

Nice to have

• Experience developing high-performance math libraries, including GEMM, convolution, reduction, or FFT kernels.
• GPU assembly-level tuning or compiler optimization experience.
• Familiarity with OpenMP or oneTBB.
• Basic understanding of deep learning primitives and upstream framework usage.

Culture & Benefits

• Work on a high-impact open-source library used across millions of devices.
• Collaborate with experts in GPU compilers, hardware architecture, and performance libraries.
• Access and influence next-generation discrete GPU software stacks.
• Competitive pay, stock programs, quarterly bonuses, healthcare, retirement benefits, and vacation.
• Hybrid flexibility with time split between the assigned site and off-site work.