Software Engineer - ML Kernels & Runtime

At Graphcore, we’re building the future of AI compute. We’re a team of semiconductor, software and AI experts, with deep experience in creating the complete AI compute stack - from silicon and software to infrastructure at datacenter scale. As part of the SoftBank Group, backed by significant long-term investment, we are delivering key technology into the fast-growing SoftBank AI ecosystem. Graphcore is expanding its teams around the world to meet the vast and exciting AI opportunity.

Essential:

• Excellent programming and scripting skills using C++ and Python
• Ability to solve challenging problems creatively, think outside the box, and see the bigger picture beyond the specific software component
• Excellent written and oral communication skills, good work ethics, and a high sense of teamwork
• Commitment to producing quality work and being a team player

Desirable:

• Strong understanding of processor architectures
• Experience or solid understanding of algorithmic performance including vectorisation, memory hierarchy, threading, lock-free patterns, and profiling on Linux
• Knowledge of at least one BLAS/DNN stack and ability to read/extend kernels
• Awareness of CPU micro-optimisations and numerical stability/trade-offs across FP32/FP16/BF16

• Implement kernels for linear algebra and tensor ops (GEMM, batched GEMM, convolutions, reductions, elementwise and fused operations) in C++
• Contribute towards performance and correctness by adding microbenchmarks, regression tests, and numerics validation
• Profile and optimise for next generation of AI hardware focusing on threading, cache locality, memory layout, and kernel launch efficiency
• Debug issues, resolve bugs and generally improve the quality and functionality of the product
• Actively engage in and support Agile ways of working within the team