Senior Software Engineer – AI Middleware
Optimize collective communication libraries including NCCL and RCCL across Cornelis Networks interconnect fabrics.
About the role
This role is responsible for designing, developing, and upstream-enabling Cornelis Networks' AI communication middleware for distributed AI workloads. The focus centers on enabling and optimizing collective communication libraries such as NCCL/RCCL over Cornelis Networks' interconnects.
What you'll do
- Design and implement performance-critical features for collective communication library enablement on Cornelis Networks' fabrics.
- Optimize distributed training performance across multi-node, multi-GPU configurations, including GPU-direct transfers, IPC, and CPU/GPU synchronization.
- Profile distributed AI workloads, identify bottlenecks across software and hardware stacks, and tune frameworks such as PyTorch Distributed, TensorFlow/XLA, JAX, DeepSpeed, and Megatron-LM.
- Develop benchmarks and microbenchmarks aligned with real model performance, and contribute upstream to AI communication and distributed training projects.
- Prototype and validate Ultra Ethernet capabilities for AI collective communication.
- Collaborate with kernel/driver, switch, performance, and systems teams to analyze traces and resolve advanced escalations.
What you'll need
- 8+ years of experience in high-performance systems programming in C/C++ on Linux.
- Strong experience with GPU communication stacks including CUDA/ROCm and NCCL/RCCL.
- Ability to optimize distributed training performance using profiling and tracing.
- Understanding of collective communication concepts and topology awareness.
- Experience delivering production-quality code and making open-source contributions in relevant areas.
Nice to have
- Experience with AI frameworks such as PyTorch Distributed, DeepSpeed, and Megatron-LM.
- Familiarity with libfabric/OFI, UCX, and RDMA concepts.
- Experience with RoCEv2 and Ultra Ethernet.
- Experience building cluster-scale performance test infrastructure.
Skills: AI communication middleware, distributed AI workloads, NCCL/RCCL, distributed training performance, CUDA/ROCm
This role has been open 144 days — well above the 53-day median for Infrastructure/Platform Software roles.
Infrastructure/Platform Software · GPU Software Stack
|
Open roles in category
310
|
Median days open
53 d
|
Category median salary (all employers)
$221k
|
See the full market breakdown ▾Category comparison, skills in demand, and who else is hiring
| Metric | Cornelis Networks | All employers we track in this specialty (310 roles · 23 employers) |
|---|---|---|
| Open roles in this specialty | 1 | 310 |
| Open roles in the wider Software, Firmware & Systems family | 9 | 6531 · 156 employers |
| Median days open | 144 d | 53 d (+91 d vs this employer) |
| Median salary (USD postings) | — | $221k |
Skills observed across this category: AI communication middleware, distributed AI workloads, NCCL/RCCL, distributed training performance, CUDA/ROCm
Who's hiring in this category
- NVIDIA · 160 open roles · median 68 d
- AMD · 63 open roles · median 57 d
- Intel Corporation · 32 open roles · median 21 d
- Qualcomm · 17 open roles · median 198 d
- Arm Holdings · 9 open roles · median 11 d
- Graphcore · 4 open roles · median 18 d
How we counted: 310 open Infrastructure/Platform Software (GPU Software Stack) roles from 23 employers tracked in the SemiconductorJobs index, counted 8 Nov 2026. Specialty figures count only roles carrying this exact specialty label, so an employer's related work in neighbouring specialties is not included there — it is counted in the wider Software, Firmware & Systems family row. Figures refresh nightly.