Senior Performance Engineer | AI Infrastructure | Cambridge (Hybrid) |

b Overview /b p In this role you will quantify and model the real-world performance of AI training and inference to answer whether optimisations or different accelerators pay off. You’ll sit between research and engineering to translate live data into actionable insights that influence buying decisions and system design. You’ll work on models and calculators that forecast hardware efficiency at scale, contributing to a mission-driven, well-backed organisation accelerating AI progress. This is a rare opportunity to shape infrastructure decisions that benefit the broader AI ecosystem. /p b Pay / Benefits /b ul li Competitive salary /li li pension /li li hybrid from Cambridge office /li /ul b Responsibilities /b ul li Bridge research and engineering to extract live training/inference metrics /li li Build performance models and calculators to forecast system behavior and cost/benefit of optimisations /li li Assess whether architecture changes or accelerators deliver ROI and inform procurement and system design /li li Shape decisions that affect the organisation and its members through data-driven insights /li /ul b Key requirements /b ul li Degree in computer science, mathematics, or adjacent field /li li Experience building performance models or calculators that forecast system behavior /li li Hands-on GPU/accelerator code optimisation (CUDA or similar) /li li Strong understanding of how LLMs and DL models run on hardware (training vs inference, matrix multiplication, KV-caching) /li li Familiarity with profiling tools (Nsight, PyTorch Profiler) and monitoring stacks (Prometheus, Grafana) /li li Python for data work (Pandas, NumPy) and scripting /li /ul ul li collaboration across research, engineering, and operations /li li analytical mindset with attention to detail /li li ability to translate data findings into actionable decisions /li li CUDA or similar GPU/accelerator optimisation /li li Nsight /li li PyTorch Profiler /li /ul