AI – CPU/NPU Performance Engineer

  • Full Time
  • Cambridge
  • Location: Cambridge
  • Salary: £70000 - £100000 per annum + standard benefits, bonus

Website McGregor Boyall Associates Limited

McGregor Boyall is a privately owned global recruitment consultancy founded in 1987. We are headquartered in the City of London, with additional offices covering the UK & Europe, the Middle East and North America. We provide permanent, contract and project-based recruitment services focusing on the mid-senior candidate market. We deliver within Financial Services, Commerce & Industry, and the Public Sector. Our primary specialisms cover all technology verticals and core business functions.

For a leading telecomunications firm I am seeking a highly experienced performance engineer to join a workload modelling team for AI workloads.

You will be responsible for designing and implementing tools and methodologies for workload modeling and simulation, as well as conducting architectural studies and software/hardware co-optimization for next-generation processors and accelerators. You should have extensive experience in workload modelling, CPU/NPU architecture, and performance analysis, and be proficient in various tools and frameworks.

Key responsibilities:

  • Design and implement tools for workload modelling and simulation
  • Conduct architectural studies and software/hardware co-optimization
  • Analyze and extract features of real-world AI workloads
  • Develop strategies for next-generation processor features


  • Extensive experience in workload modelling and CPU/NPU architecture
  • Proficient in performance analysis tools and simulators
  • Strong programming skills in C++ and Python
  • Experience with deep learning frameworks
  • Excellent collaboration and interpersonal skills


  • Experience with QEMU, DynamoRIO, and performance simulators
  • Experience with CUDA or OpenCL programming

McGregor Boyall is an equal opportunity employer and do not discriminate on any grounds.

Upload your CV/resume or any other relevant file. Max. file size: 18 MB.