Performance Engineer

CommonAI CIC; Cambridge; On-site; GBP 65,000 - 90,000; Full time

Benefits offered by this job

Collaborative environment

High impact in growing org

Competitive salary and pension

Professional development opportunities

Networking with tech & academia

Office near Cambridge train station

Job summary

CommonAI CIC is seeking a Performance Engineer to join a growing team. You will work with AI researchers and software engineers to understand and model performance across inference and training workloads, instrumenting metrics and building predictive models.

You will use these models to forecast the impact of software optimisations and architectural or hardware changes, guiding cost-effective performance improvements and hardware choices for our in-house and member projects.

Qualifications

• Experience building mathematical models and forecasts of system behavior.
• Experience optimising code on GPUs/accelerators (e.g. CUDA).
• Solid understanding of computer architecture and DL inference vs training.
• Proficiency with profiling/monitoring tools (Nsight, PyTorch Profiler, Prometheus, Grafana).
• Python for data analysis (Pandas, NumPy) and scripting.

Responsibilities

• Build and instrument granular performance metrics for inference/training workloads.
• Develop models to predict impact of optimisations on performance and cost.
• Influence hardware purchasing decisions and architectural optimisation based on data.

Skills

Mathematical modeling

CUDA optimization

Python data analysis

Profiling expertise

Performance forecasting with Excel/SS

Tools

NVIDIA Nsight

PyTorch Profiler

Prometheus

Grafana

Job description

CommonAI CIC is a non-profit membership organisation, founded on a belief in collaborative engineering for the safe and responsible development of foundational AI technologies. A place where AI startups, enterprises large and small, public sector bodies and academia can share resources and knowledge, to codevelop and grow businesses, fast.

We are seeking a Performance Engineer to join our rapidly growing team. In this role, you will work with AI researchers and software engineers to build up a detailed understanding of how their applications are performing. You will instrument and collect granular metrics from inference and training jobs and use that information to develop sophisticated mathematical models that predict how software optimisations and architectural or hardware changes will impact system performance.

Your work will directly influence both our in-house and member's hardware purchasing decisions and architectural optimisations, ensuring teams can run AI workloads efficiently and cost-effectively.

Requirements

• Experience building insightful mathematical models and performance calculators (Excel/Google Sheets or Python modeling experience) to forecast system behavior.
• Optimisation of code running on GPUs and/or other accelerators (e.g. CUDA).
• Solid understanding of computer architecture fundamentals and how LLMs and Deep Learning models execute on that hardware (inference vs. training, matrix multiplication, KV-caching, etc.).
• Proficiency with profiling tools (NVIDIA Nsight, PyTorch Profiler) and monitoring stacks (Prometheus, Grafana).
• Capability to work in Python for data analysis (Pandas, NumPy) and scripting.

The Following Are Also Highly Valued

• Post-graduate degrees and research experience in relevant fields (please list your publications).
• Deep understanding of inference serving frameworks (e.g. vLLM).
• Background in statistical analysis.
• Contributions to open source and/or research projects.

Benefits

• A collaborative and supportive work environment
• The opportunity to have a high impact in a growing organisation
• Competitive salary package and pension
• Professional development opportunities
• Networking opportunities with influential people from across the tech sector and academia
• A vibrant office environment located a few minutes' walk away from Cambridge train station