Senior Software Engineer — cuEquivariance
Core
Building and optimizing CUDA kernels and Python interfaces for geometric neural networks to accelerate biological discovery and scientific research.
Role type
Senior IC GPU software engineer (geometric deep learning)
Builds
Production GPU kernels and software libraries for equivariant deep learning
Domain
Scientific computing, GPU computing, geometric machine learning
Deliverable
production ML models
Required skills
CUDA kernel engineering, C++, Python, GPU memory hierarchy, performance profiling, geometric machine learning concepts
Preferred skills
Triton kernel development, mixed-precision algorithm design, open-source contributions to geometric ML
Technologies
CUDA, PyTorch, JAX, Triton
Responsibilities
Build and optimize CUDA kernels for equivariant neural network primitives; deliver end-to-end GPU-accelerated geometric ML primitives; build and maintain PyTorch and JAX interfaces; drive CI/CD infrastructure for multi-GPU builds; collaborate with research teams to translate prototypes into production kernels; engage with third-party framework developers to ensure integration.
Seniority
Senior, hands-on IC
