GPU HPC Middleware Engineer

Intel; Austin (TX); On-site; USD 111,000 - 211,000; Full time

Benefits offered by this job

Stock bonuses

Health benefits

Vacation

Job summary

Intel's Communication Runtimes team is seeking a Middleware Development Engineer to optimize software communication libraries for HPC and AI workloads at scale. You will identify bottlenecks in the oneCCL library running large-scale AI workloads and help design performance improvements.

You will collaborate with teams working on SHMEM, MPI, and GPUs to maximize low latency, high bandwidth, and reliability across Intel architectures. This on-site role supports climate modeling and AI systems.

Qualifications

• Bachelors/Master/PhD with 3+ years / 1+ years / 3 months software development experience
• Experience in distributed computing systems, HPC libraries (MPI/SHMEM/oneCCL/NCCL)
• Proficiency in C and C++ programming
• Experience in Linux environments
• Experience multithreaded programming
• Strong analytical skills and problem solving

Responsibilities

• Identify performance bottlenecks and features for Argonne AI COE workloads
• Optimize distributed runtime software for low latency and high bandwidth
• Collaborate with cross-functional teams to define specs
• Troubleshoot issues across hardware/software stack layers
• Contribute to innovations for HPC/AI at scale
• Partner with engineering/architecture to maximize performance

Job description

Intel's Communication Runtimes team is seeking a Middleware Development Engineer to optimize software communication libraries for HPC and AI workloads at scale. You will identify bottlenecks in the oneCCL library running large-scale AI workloads and help design performance improvements.

You will collaborate with teams working on SHMEM, MPI, and GPUs to maximize low latency, high bandwidth, and reliability across Intel architectures. This on-site role supports climate modeling and AI systems.