Pytorch Engineer

Graphcore is one of the world’s leading innovators in Artificial Intelligence compute. It develops hardware, software, and systems infrastructure to enable the next generation of AI breakthroughs and power AI adoption across industries. As part of the SoftBank Group, Graphcore is a member of a family of companies focused on enabling Artificial Super Intelligence. The Frameworks team ensures Graphcore hardware works seamlessly with popular machine learning tools such as PyTorch, Triton, Jax, and TensorFlow, collaborating closely with open-source communities and other internal teams.

Essential:

• Strong software engineering skills.
• Experience in Python development.
• Experience developing performant C++ applications in a commercial setting.
• At least one of the following: experience writing ML kernels, compiler development, ML frameworks development or usage, or experience in a computationally intensive engineering field.

Desirable:

• Knowledge of AI/ML.
• Understanding of computing architectures.
• Experience maintaining and supporting complex code bases.
• Experience profiling and optimizing high-performance code.

• Designing and implementing new features and maintaining existing ones.
• Developing and maintaining unit, component, and integration tests.
• Optimising the software stack for performance.
• Managing complex technical tasks with cross-team dependencies.
• Managing code quality, complexity, and technical debt.
• Providing technical leadership and contributing to the team's technical direction and processes.
• Contributing to documentation including user manuals and tutorials.
• Conducting code reviews.
• Resolving regressions, performance issues, and software defects.
• Coaching and mentoring team members.
• Contributing to a collaborative team culture.
• Driving continuous improvements in ways of working.
• Working with upstream open-source development teams.