NPU Systems Performance Analysis Engineer

Full Job Description

NPU Systems Performance Analysis Engineer

I am seeking someone experienced in Systems Performance Analysis, specifically using performance simulators to develop benchmarks and to analyse and identify bottlenecks in NPUs / AI Accelerators. Must be able to work closely with HW and SW teams to provide optimisation recommendations by understanding performance opportunities on current processor architecture. Based in Cambridge on a hybrid basis.

Primary Responsibilities

Using ML models to carry out performance analysis on NPUs.

Working closely with the modelling and performance teams utilising the simulator for NPU performance analysis.

Pinpointing problems in the HW / problems in the compiler and proposing optimisations, bottlenecks, latency etc.

Required Experience

Masters or PhD in Computer Science, Electronic Engineering or similar.

At least 4 years of industry experience working on system performance analysis for NPUs / GPUs.

Understanding of NPU (neural processing unit) hardware and software.

Strong knowledge of CNN and Transformer Architecture.

Experience with ML frameworks like Pytorch/Tensorflow and programming languages C++ and Python.

Desired Experience

Industry experience at NPU, GPU, semiconductor manufacturers.

What’s on offer

Competitive Salary (£80k-120k)

Hybrid working (1-2 day’s onsite minimum, preference for more on-site if local)

Annual Bonus

25 days holiday

And more!