Senior Math Libraries Engineer - Direct Sparse Solvers
Core
Designing, implementing, and optimizing direct sparse solvers for GPU architectures to accelerate scientific and engineering simulations, data analytics, and AI.
Role type
Senior IC high-performance computing engineer (GPU kernels)
Builds
cuDSS, a CUDA library for direct solvers for sparse linear systems
Domain
High-performance computing / Scientific computing / GPU acceleration
Deliverable
production ML models
Required skills
C++, CUDA, parallel programming, sparse linear algebra primitives, numerical analysis, floating-point arithmetic, GPU hardware architecture, low-level performance optimization
Preferred skills
Direct solver techniques (reordering, multi-frontal factorizations), CI/CD systems, large-scale computing technologies (PDE solvers, eigenvalue solvers)
Responsibilities
Designing and optimizing direct sparse solvers for existing and future GPU architectures; collaborating with library engineers, QA, and interns on development, testing, and release; working with product management to define technical roadmaps; improving library quality and maintainability through re-architecting
Seniority
Senior, hands-on IC
