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

Overview

In this role you will work with senior engineers to turn live training and inference metrics into predictive models and calculators. You help assess whether optimisations or new hardware setups are worthwhile before spending, influencing cost and performance. You gain hands-on experience with GPU workloads, hardware-aware modelling, and data tools in a collaborative, research-oriented environment. This is an early-career opportunity at a Cambridge-based nonprofit aiming to reduce duplicated AI infrastructure effort across academia and industry.

Pay / Benefits

• pension
• hybrid work model
• Cambridge office
• exposure to wider AI and academic community

Responsibilities

• collect real metrics from live training/inference
• build performance models and calculators to answer optimization questions
• assess cost/perf trade-offs of different setups
• take ownership with senior support
• work closely with senior engineers on real problems

Key requirements

• postgraduate research background
• PhD ideally or exceptional master’s with relevant coursework
• strong grasp of computer architecture and how LLMs run on hardware
• experience building performance models or forecasting tools
• hands-on GPU/accelerator code experience (CUDA or similar)
• familiarity with profiling tools (Nsight, PyTorch Profiler) and monitoring stacks (Prometheus, Grafana)
• strong Python for data work (Pandas, NumPy)
• scripting ability

• strong analytical thinking
• collaborative mindset
• clear communication with engineers and researchers
• CUDA or similar GPU/accelerator programming
• Nsight
• PyTorch Profiler

…