ML Frameworks Engineer for AI Hardware (Triton/PyTorch)

Graphcore; Poland; Hybrid; PLN 150,000 - 210,000; Full time

Benefits offered by this job

Flexible working

Healthcare and dental

Phantom equity

Gym access

Life insurance

Retirement savings

EAP

Job summary

Graphcore is seeking an engineer to build and support framework software that helps AI developers unlock Graphcore hardware. You will enable accelerators to work with ML frameworks like PyTorch and Triton, design, implement and optimize critical software, and deliver features with strong ownership.

Joining the Triton team, you will review code, write technical documentation and collaborate across teams and open-source communities.

Qualifications

• Strong software engineering skills with the ability to manage quality, complexity and technical debt.
• Experience developing in Python and C++.
• Experience in compiler development.
• Ability to manage complex technical tasks with cross‑team dependencies.
• Clear communication skills with a proactive approach to collaboration and technical leadership.
• Interest in AI, ML frameworks, performance optimisation or computationally intensive engineering.

Responsibilities

• Design, implement, optimise and support critical software.
• Help Graphcore accelerators work seamlessly with state-of-the-art ML frameworks.
• Deliver new features, review code, write technical documentation and work with upstream communities.
• Coordinate open-ended technical work across teams.

Skills

Software engineering

Python

C++

Compiler development

Cross-team coordination

Communication

AI/ML frameworks interest

Job description

Graphcore is seeking an engineer to build and support framework software that helps AI developers unlock Graphcore hardware. You will enable accelerators to work with ML frameworks like PyTorch and Triton, design, implement and optimize critical software, and deliver features with strong ownership.

Joining the Triton team, you will review code, write technical documentation and collaborate across teams and open-source communities.