Senior AI Inference & Kernel Engineer

Intel; Austin (TX); Hybrid; USD 189,000 - 315,000; Full time

Benefits offered by this job

Stock bonuses

Health benefits

Vacation

Job summary

Intel is seeking an AI Infrastructure Engineer to push LLM inference on Intel GPUs to new limits in Santa Clara and across multiple US locations. You will profile bottlenecks, write high-performance GPU kernels, and upstream improvements to open-source projects like vLLM and PyTorch.

You will work end-to-end across the stack, shape hardware roadmaps with roofline analysis, and collaborate with architecture and compiler teams to optimize performance for state-of-the-art generative AI workloads.

Qualifications

• Experience in GPU computing and AI systems development.
• Strong proficiency in C++ and Python; reading/modifying complex code.
• Higher degree or equivalent experience is preferred.

Responsibilities

• Own the end-to-end optimization pipeline for running state-of-the-art LLMs on Intel GPUs.
• Profile, diagnose, and resolve cross-stack performance bottlenecks.
• Design and optimize custom high-performance kernels for attention, MoE, quantization, and fusions.
• Upstream architectural improvements into open-source projects like vLLM, SGLang, PyTorch.
• Collaborate with architecture and compiler teams to shape GPU roadmaps based on GenAI workloads.
• Demonstrate passion for AI infrastructure and performance optimization.

Skills

GPU computing

C++

Python

HPC

Education

Bachelor's degree in Computer Science or related field

Tools

Triton

CUDA/CUTLASS

SYCL

Job description

Intel is seeking an AI Infrastructure Engineer to push LLM inference on Intel GPUs to new limits in Santa Clara and across multiple US locations. You will profile bottlenecks, write high-performance GPU kernels, and upstream improvements to open-source projects like vLLM and PyTorch.

You will work end-to-end across the stack, shape hardware roadmaps with roofline analysis, and collaborate with architecture and compiler teams to optimize performance for state-of-the-art generative AI workloads.