Remote AI/HPC Middleware Engineer (MPI/CCL)

Cornelis Networks; New York (NY); On-site; USD 120,000 - 180,000; Full time

Benefits offered by this job

Equity

Health benefits

401(k) with company match

Open Time Off

Job summary

Cornelis Networks is seeking a Software Engineer to help develop and optimize AI/HPC middleware. You will contribute to enabling and tuning MPI/SHMEM and AI stacks (NCCL/RCCL and related libraries) over our interconnects, with guidance from senior engineers.

This is a remote US role with a strong total rewards package including equity. You will gain exposure across the stack—drivers, kernel, switches, systems, and application frameworks—while delivering features, testing, and upstream

Qualifications

• 2+ years of systems programming in C/C++ on Linux.
• Experience with HPC middleware (MPI/SHMEM) or AI collectives (NCCL/RCCL).
• Familiarity with profiling/tracing tools and performance analysis.
• Foundational understanding of networking/RDMA concepts.

Responsibilities

• Enable and optimize HPC middleware (MPI/SHMEM) and AI middleware (NCCL/RCCL).
• Integrate middleware with transports and provider layers (libfabric/OFI, UCX).
• Contribute upstream to MPI/SHMEM projects and CCL ecosystems.
• Collaborate across kernel/driver and switch teams for end-to-end performance.
• Analyze performance traces and reproduce customer issues.

Skills

C/C++

Linux

Profiling tools

System performance analysis

HPC middleware knowledge

Education

Bachelor's degree in Computer Science or Engineering

Master's or Doctorate a plus

Tools

MPI/SHMEM

NCCL/RCCL

CUDA/ROCm

libfabric

UCX

Job description

Cornelis Networks is seeking a Software Engineer to help develop and optimize AI/HPC middleware. You will contribute to enabling and tuning MPI/SHMEM and AI stacks (NCCL/RCCL and related libraries) over our interconnects, with guidance from senior engineers.

This is a remote US role with a strong total rewards package including equity. You will gain exposure across the stack—drivers, kernel, switches, systems, and application frameworks—while delivering features, testing, and upstream