Staff Software Engineer - AI Compute Libraries & Performance
Core
Design and implement high-performance AI compute kernels (linear algebra, tensor ops) for next-generation AI hardware, ensuring numerical correctness and low-level performance optimization.
Role type
Staff Software Engineer (AI Compute Libraries)
Builds
High-performance compute libraries (ML Kernels & Runtime) for Graphcore's IPU hardware
Domain
Semiconductor / AI Infrastructure / High-Performance Computing
Deliverable
production ML models
Required skills
C++, linear algebra, tensor operations, numerical validation, performance profiling, low-level optimization, threading, cache locality, memory layout, kernel launch efficiency, debugging
Technologies
C++, GEMM, convolutions, reductions, elementwise operations
Responsibilities
Design and implement kernels for linear algebra and tensor ops; Provide technical leadership and guide complex engineering decisions; Own performance and correctness via microbenchmarks and regression tests; Profile and optimize for next-gen AI hardware; Debug issues and improve product quality
Seniority
Staff, hands-on IC with technical leadership
