Fellow Software Development Engineer, GEMM Optimization
Core
Architect and develop features in the tensile-lite assembly code generator for General Matrix Multiplication (GEMM) and GEMM + X to deliver best-in-class matrix-multiply performance across AMD GPU hardware.
Role type
Fellow Software Development Engineer (GPU Microarchitecture & Code Generation)
Builds
High-performance GEMM kernels and assembly code generators for AMD GPUs
Domain
Semiconductor / GPU Architecture / High-Performance Computing
Deliverable
production ML models
Required skills
GPU microarchitecture, GEMM algorithms, C++, assembly-level programming, parallel computing, memory hierarchy optimization, GPU compiler backends
Preferred skills
AMD GPU architectures (CDNA/RDNA), NVIDIA Hopper/Blackwell experience, hardware-software co-design, performance profiling, algorithm innovation
Technologies
C++, Assembly, Python, Tensile-lite, CDNA, RDNA, Hopper, Blackwell
Responsibilities
Analyze GPU hardware specifications and align kernel design with silicon capabilities; Develop support for new ISA and HW/SW optimization features in the code-generator; Innovate and implement new algorithms for GEMM and GEMM + X; Profile and root-cause GEMM performance bottlenecks; Partner with hardware architects and customers to deliver targeted performance improvements
Seniority
Fellow, strategic IC with deep hardware expertise
