Senior Software Engineer — cuEquivariance - CA, Santa Clara, United States
$184,000 - $356,500 USD yearly
Originally posted 22 May 2026 by the employer. Relisted by the employer on 14 August 2026.
Work on the cuEquivariance library implementing CUDA kernels and PyTorch/JAX interfaces for geometric neural networks.
About the role
The work spans CUDA kernel engineering, Python library development involving PyTorch and JAX, and direct collaboration with research teams and external framework developers.
What you'll do
- Build, implement, and optimize CUDA kernels for equivariant neural network primitives—including tensor products, segmented polynomials, and triangle-based operations—targeting peak performance across NVIDIA GPU generations. - Build and maintain interfaces for PyTorch and JAX that expose cuEquivariance primitives to application developers and researchers. - Drive CI/CD infrastructure for multi-GPU kernel builds, automated correctness testing, and performance regression tracking. - Collaborate with Applied Science and research teams to evaluate new equivariant architectures and translate prototypes into production kernels. - Engage directly with third-party framework developers and partners to align on interfaces and ensure software integrates cleanly into production pipelines.
What you'll need
- 6+ years of software engineering experience with a strong background in CUDA and GPU programming. - Deep proficiency in C++ and Python, with experience building and shipping production libraries used by external developers. - Solid foundation in GPU computing, including memory hierarchy, warp-level execution, occupancy, and performance profiling methodology. - Experience building or contributing to production scientific software libraries, ML frameworks, or developer-facing GPU APIs. - Familiarity with concepts in geometric machine learning—such as equivariance, group representations, irreducible representations, or tensor products. - BS/MS in Computer Science, Physics, Applied Mathematics, or a related field (or equivalent experience).
Skills: CUDA kernels, GPU-accelerated geometric ML primitives, PyTorch and JAX interfaces, CI/CD infrastructure, equivariant neural network
This role has been open 154 days — well above the 55-day median for Infrastructure/Platform Software roles.
Infrastructure/Platform Software · GPU Software Stack
|
Open roles in category
242
|
Median days open
55 d
|
Median salary
$220k
|
See the full market breakdown ▾Category comparison, skills in demand, and who else is hiring
| Metric | NVIDIA | All employers we track in this specialty (242 roles · 20 employers) |
|---|---|---|
| Open roles in this specialty | 122 | 242 |
| Open roles in the wider Software, Firmware & Systems family | 997 | 6277 · 144 employers |
| Median days open | 57 d | 55 d (+2 d vs this employer) |
| Median salary (USD postings) | — | $220k |
Skills observed across this category: CUDA kernels, GPU-accelerated geometric ML primitives, PyTorch and JAX interfaces, CI/CD infrastructure, equivariant neural network
Who's hiring in this category
- NVIDIA (this employer) · 120 open roles · median 57 d
- AMD · 52 open roles · median 53 d
- Intel Corporation · 25 open roles · median 14 d
- Qualcomm · 18 open roles · median 128 d
- Arm Holdings · 5 open roles · median 121 d
- Graphcore · 4 open roles · median 3 d
How we counted: 242 open Infrastructure/Platform Software (GPU Software Stack) roles from 20 employers tracked in the SemiconductorJobs index, counted 24 Oct 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.