PyTorch Engineer

PyTorch Engineer Overview

This role sits within Graphcore’s Frameworks organisation on the PyTorch team, where the focus is on making Graphcore hardware feel like a natural extension of the machine learning tools engineers already use. The work is centred on building and improving the software layer that connects Graphcore accelerators with modern ML frameworks, so that researchers and engineers can get better compatibility, stronger performance and a smoother development experience without changing the way they build models.

As a Software Engineer on the PyTorch team, you will work across PyTorch, Triton and related technologies. The role involves designing, implementing, optimising and supporting features that help Graphcore hardware integrate cleanly into the wider ML ecosystem. You will also contribute upstream to PyTorch, maintain and extend the team’s hybrid Python and C++ extension layer, and help bring compiler backend capabilities into torch.compile. The position is hands-on and technical, with a strong emphasis on practical engineering, open-source collaboration and visible impact.

What you will do

• Develop software that makes Graphcore accelerators work naturally with the frameworks and workflows ML engineers already rely on.
• Design, build, tune and maintain functionality across PyTorch, Triton and adjacent technologies.
• Submit improvements upstream to the PyTorch project and work within a large open-source codebase.
• Support the team’s mixed Python/C++ extension layer and improve how it integrates with the rest of the stack.
• Help connect compiler backend features with torch.compile so users can access Graphcore capabilities more easily.
• Improve framework compatibility, runtime efficiency and the overall developer experience for ML engineers and researchers.
• Work in a SCRUM team that delivers features, reviews code, documents work and supports users.
• Collaborate with colleagues in software, AI/ML, research and hardware to make sound technical decisions.
• Investigate unfamiliar systems, solve engineering problems and contribute to the health of the codebase as you learn.
• Write maintainable code and thoughtful designs that support long-term quality and reliability.

What the team is like

You will join the PyTorch team inside Frameworks. The group works in a fast-moving but careful way, balancing speed with attention to user needs and long-term code quality. The culture values clear ownership, practical decision-making and close communication across disciplines. Engineers are trusted to explore, ask questions, speak up and improve the system as they gain understanding.

The team is a good fit for people who enjoy open-source communities, technically challenging problems and work where the results are visible. There is an expectation that you will move quickly, but with care for maintainability and the experience of the people using the software.

What they are looking for

• Strong general software engineering skills and a collaborative mindset.
• Experience building Python libraries.
• Experience creating performant C++ applications in a commercial setting.
• Some understanding of linear algebra, floating-point arithmetic or numerical analysis.
• Confidence using Git and taking part in code review in a professional engineering environment.
• Willingness to explore large, unfamiliar open-source codebases on your own and with others.
• Curiosity, adaptability and the ability to learn quickly in a complex technical environment.
• Interest in open-source software, deep technical work and engineering that has a clear impact on users.

Benefits and perks

• Flexible working to help balance professional and personal commitments.
• Generous leave to support rest and time away from work.
• Pension support with employer matching of up to 5%.
• Phantom equity so you can share in the company’s success.
• Office spaces designed for collaboration, with free food and an on-site barista.
• Income protection and life assurance for financial peace of mind.
• A flexible benefits scheme with options such as private medical insurance and dental cover.
• Optional extras including dental cover, a health cash plan, private medical insurance, a cycle-to-work scheme and give-as-you-earn.

Inclusion and hiring process

Graphcore says it welcomes applicants from all backgrounds and experiences and is committed to building an inclusive environment where everyone can do their best work. The company describes itself as an equal opportunity employer and says it values the different strengths and perspectives people bring. If you need adjustments during the interview process, you can let them know and they will support you.

The application form asks for a CV and cover letter, along with basic contact details and a full legal name including any middle names. It also asks candidates to confirm whether they have the legal right to work in the UK and to select their right-to-work status. Applicants may also provide a LinkedIn profile, website and referral source. The form includes optional confidential diversity questions, which Graphcore says are anonymised and used only for statistical purposes to improve hiring inclusivity.

About Graphcore

Graphcore is presented as a major innovator in AI compute, building hardware, software and systems infrastructure intended to support the next wave of AI progress and broader adoption of AI across industries. As part of the SoftBank Group, it is described as being connected to a wider family of companies behind transformative technologies and a shared ambition to make advanced AI benefits widely accessible. The company brings together AI researchers, silicon designers, software engineers and systems architects to solve complex problems and deliver progress in AI compute.

If you want to help shape how PyTorch runs on Graphcore hardware, the company invites you to apply and join the team.

Degree Requirement: Not Specified

Visa Sponsorship May be