SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: PLN 350,000 - 450,000 / annual
Graphcore is seeking a Staff Software Engineer to own performance insights across its proprietary AI hardware and software stack. In this role, you will shape end-to-end performance by connecting model execution, compiler behaviour, runtime choices and hardware utilisation. You will find bottlenecks, explain regressions and guide improvements across component boundaries, helping teams make local changes that improve the whole system.
Your responsibilities include building benchmarks, profiling large-scale ML workloads and translating low-level signals into actionable decisions for engineers. You will work with framework, compiler, runtime, infrastructure and product teams to validate real impact. This role offers rare scope across the full ML stack, from single chip behaviour to distributed systems.
The ML Software Performance Analysis team works across the ML software organisation, using benchmarks, profiling and system analysis to expose performance truth. Ownership is broad—you will publish reports, challenge assumptions and follow issues across frameworks, compilers, runtime and infrastructure until impact is proven. Decisions are made through evidence, and the team values engineers who think big, act fast, take responsibility, speak up and lead beyond their own area.
Graphcore is one of the world's leading innovators in AI compute, developing hardware, software and systems infrastructure to unlock the next generation of AI breakthroughs. As part of the SoftBank Group, Graphcore is a member of an elite family of companies responsible for transformative technologies.
Requirements:
• Strong Python, C or C++ skills in performance-sensitive software
• Deep understanding of computer architecture, profiling and low-level system behaviour
• Experience benchmarking and analysing complex distributed systems
• Ability to translate performance data into clear engineering recommendations
• Working knowledge of Linux development environments and tooling
• Familiarity with ML frameworks, hardware acceleration or collective communication
The company welcomes transferable skills and diverse experiences, and also welcomes engineers returning to the profession after a career break.
