Senior Software Engineer - AI Compute Libraries & Performance

2 weeks agoSenior Bristol, UK Ai Jobs by Graphcore

Skills

Benchmarking Blas C#Convolution Dnn Gemm Linear Algebra Linux Memory Hierarchy Numerical Stability Processor Architecture Profiling Python Pytorch Regression Testing Tensor Operation Threading Vectorisation

Candidate Availability

Bristol · Required United Kingdom · Required

Required and preferred rules are kept separate and reflect the wording in the original posting.

About the Role

You will design and implement high-performance C++ kernels for linear algebra and tensor operations, including GEMM, convolutions, reductions and fused operations. You will benchmark, test, profile and optimize kernels across AI hardware, improve numerical reliability, debug issues, and mentor colleagues.

Requirements

Excellent C++ and Python programming and scripting skills
Understanding of processor architectures and profiling on Linux
Strong written and oral communication skills
Knowledge of vectorisation, memory hierarchy, threading and lock-free patterns
Experience with a BLAS or DNN stack
Experience with CPU micro-optimisation and numerical stability across FP32, FP16, BF16 and FP8
Experience integrating native code into PyTorch or similar frameworks
Knowledge of ABI/API stability and Linux packaging

Responsibilities

Design and implement C++ kernels for linear algebra and tensor operations
Own performance and correctness through microbenchmarks, regression tests and numerical validation
Profile and optimize threading, cache locality, memory layout and kernel launch efficiency
Debug issues and improve product quality and functionality
Support Agile ways of working
Mentor colleagues and provide technical guidance

Benefits

Unlimited annual leave
Up to 5% matched pension
Phantom equity
Flexible work location and arrangements
Free food
On-site barista
Health cash plan
Income protection
Life assurance
Private medical insurance
Dental plan