Senior Software Engineer, CUTLASS Platform
Core
Designing abstractions for Tensor Core and GPU hardware features in MLIR, Python, and C++ to enable high-performance linear algebra and deep learning computations.
Role type
Senior IC systems software engineer (GPU compilers & kernels)
Builds
CUTLASS open-source ecosystem for high-performance linear algebra and Tensor Core primitives
Domain
High-performance computing, GPU architecture, compiler infrastructure
Deliverable
production ML models | infrastructure
Required skills
C++ programming, software design, high-performance code generation, compiler transformations, computer architecture, parallel computing programming models
Preferred skills
MLIR dialect design, NVVM/PTX low-level kernel development, deep learning algorithms
Technologies
CUDA C++, CUTLASS Python DSL, MLIR, NVVM, PTX
Responsibilities
Develop core components including Tensor Core MMAs, copies, synchronization barriers, and schedulers; Contribute to MLIR-based backend compiler stack by designing dialects and passes; Author example kernels showcasing novel GPU hardware features; Collaborate with GPU architecture and compiler teams on programming models and performance assessment
Seniority
Senior, hands-on IC
