Workplace
On-site
Employment
Part-Time
Published
Aug 11, 2026
Closes
No date supplied
The role
About the job
Validate the ML stack that turns accelerator hardware into trusted AI performance.
This role sits where modern ML models meet Graphcore’s software and hardware stack. You will test, benchmark and validate complex systems before they reach customers.
Your work will expose regressions, correctness issues and performance limits across frameworks, models and execution environments. You will help teams understand what is working, what is breaking and why.
You will run open source models, build automated benchmarking pipelines and create targeted tests for low level ML behaviour. That includes numerical precision, quantisation, attention mechanisms, distributed execution and model subgraphs.
This is a role for someone who wants to stay close to how AI systems really work. You will not design new models, but you will make them run reliably on ambitious infrastructure.
The team and culture
The ML QA team is where Graphcore’s ML software stack comes together for validation. Work spans unit tests, full model benchmarks, distributed workloads, simulators, emulators and hardware targets.
Engineers are expected to take ownership, question assumptions and improve how testing is done. The team moves quickly, but decisions are grounded in evidence, benchmark data and technical discussion.
You will work closely with software, infrastructure and hardware teams. Squads organise around priorities, with space for engineers to shape roadmaps and raise the quality bar.
What we’re looking for
· Strong experience in Machine Learning or ML-adjacent software engineering roles.
· A solid grasp of neural networks, training, inference, numerical precision and performance trade-offs.
· Hands-on experience with PyTorch, TensorFlow, JAX, Triton or
Requirements
Department: Software Ops