Research Engineer, Infrastructure, Kernels
San Francisco
Workplace: HybridFull timeUSD 350,000 - 475,000 annuallyFunction: Research & Scientific (R&D)Education: bachelorsSkills: ["Debugging","Collaboration","Initiative","Performance profiling","Technical communication"]Design and optimize high-performance ML kernels that power large-scale language model training, including attention, matrix multiplication, gating, and normalization. Build compute primitives to reduce memory bottlenecks and improve compute efficiency, then profile performance across GPU/accelerator generations. Collaborate with research teams and systems architects to align kernel optimizations with model goals, maintain reusable kernel libraries and benchmarks, and help ensure scalability, stability, and reproducibility. Document and share technical insights via papers and talks.
Loading
Loading job details...
Preparing the role view and application actions.

