Power and Performance Engineer

ARM; Cambridge; Hybrid; GBP 45,000 - 80,000; Full time

Job summary

ARM is seeking a hands-on Power and Performance Engineer to shape profiling methodologies for pre-silicon and post-silicon platforms. You will run workloads on Linux-based devices and reference platforms, build automation harnesses, and analyze power and performance to influence architecture and product decisions.

The role requires strong scripting, data analysis, and the ability to communicate trade-offs clearly with engineering, architecture, and product teams in a hybrid work environment.

Qualifications

• Experience in pre-silicon or post-silicon power and performance analysis.
• Experience with profiling SoCs across CPU/GPU/NPU, memory, and interconnect.
• Hands-on workloads on Linux-based devices and embedded platforms.
• Ability to develop automation harnesses for profiling, data collection and analysis.
• Understanding of power, thermal and performance trade-offs in complex SoCs.
• Familiarity with profiling tools and telemetry systems.
• Strong scripting capability (Python, shell).
• Ability to translate workloads into measurable requirements.
• Excellent written and verbal communication with engineering teams.

Responsibilities

• Perform hands-on power and performance profiling of SoC platforms across CPU, GPU, NPU, memory, interconnect, and full-system workloads.
• Run benchmarks and real-world workloads on Linux-based devices, development boards, reference platforms, and pre/post-silicon environments.
• Develop and maintain automation harnesses for workload execution, profiling, telemetry, power measurement, and result analysis.
• Drive methodology for pre-silicon PnP analysis: workload selection, benchmark definition, estimation and correlation planning.
• Support post-silicon PnP analysis: platform characterization, bottleneck analysis, power/performance trade-off analysis.
• Define TDP methodology linking workloads, sustained performance, thermal constraints, and product segments.
• Define power and performance requirements for PAI products in robotics and automotive markets.
• Produce clear technical reports and recommendations for architecture, product, software, systems, and platform teams.

Skills

Power & performance analysis

SoC profiling

Linux workload profiling

Python scripting

Data analysis

Technical communication

Electronics/CS degree

Education

Degree in Electronics / Electrical / Computer Engineering or Computer Science

Tools

perf

ftrace

trace-cmd

sysfs

procfs

vendor telemetry tools

power analysers

Python

Job description

Salary: £45,000 - 80,000 per year

Requirements:

• Demonstrable experience in either pre-silicon or post-silicon power and performance analysis.
• Experience with SoC power and performance analysis across heterogeneous compute systems such as CPU, GPU, NPU, memory, or interconnect.
• Hands-on experience running workloads on Linux-based devices, development boards, embedded platforms, or pre/post-silicon environments.
• Experience using or developing automation harnesses for workload execution, data collection, profiling, and analysis.
• Understanding of power, thermal, and performance trade-offs in complex SoCs.
• Experience with profiling tools, performance counters, system traces, telemetry, or power measurement equipment.
• Strong scripting and data-analysis skills, for example Python, shell scripting, or similar tools.
• Ability to translate workload behaviour and product use cases into measurable engineering requirements.
• Strong communication skills with the ability to explain technical trade-offs to engineering, architecture, and product collaborators.
• Degree or equivalent experience in Electronics Engineering, Electrical Engineering, Computer Engineering, Computer Science, or equivalent industry experience.
• Nice to have: experience with embedded robotics development platforms.
• Nice to have: experience in platform performance analysis, silicon validation, performance modelling, systems engineering, or architecture analysis.
• Nice to have: experience with automotive, robotics, edge AI, embedded AI, or heterogeneous compute platforms.
• Nice to have: familiarity with AI/ML workloads, vision pipelines, robotics perception/planning workloads, or automotive ADAS workloads.
• Nice to have: experience defining TDP, sustained performance targets, thermal constraints, or power envelope methodologies.
• Nice to have: knowledge of Arm architecture, Arm-based SoC platforms, Linux performance tools, and embedded software ecosystems.
• Nice to have: experience with tools such as perf, ftrace, trace-cmd, sysfs, procfs, vendor telemetry tools, or power analysers.
• Nice to have: experience with benchmark automation, lab infrastructure, remote device farms, or CI-style workload execution.
• Nice to have: familiarity with pre-silicon environments such as performance models, simulators, emulation, FPGA, or virtual platforms.

Responsibilities:

• Perform hands-on power and performance profiling of SoC platforms across CPU, GPU, NPU/AI accelerator, memory, interconnect, and full-system workloads.
• Run benchmarks and real-world workloads on Linux-based devices, development boards, reference platforms, and pre/post-silicon environments.
• Develop and maintain automation harnesses for workload execution, profiling, telemetry, power measurement, and result analysis.
• Drive methodology for pre-silicon PnP analysis, including workload selection, benchmark definition, estimation, and correlation planning.
• Support post-silicon PnP analysis, including platform characterisation, bottleneck analysis, and power/performance trade-off analysis.
• Define TDP methodology for SoC products, linking workloads, sustained performance, peak power, thermal constraints, and product segments.
• Define power and performance requirements for PAI products in robotics and automotive markets.
• Produce clear technical reports and recommendations for architecture, product, software, systems, and platform teams.

Technologies:

• AI
• ARM
• Embedded
• FPGA
• Hardware
• Support
• Linux
• Python
• Robotics
• C++

More:

We are Arms Platforms and Performance Analysis team, and we are looking for a hands-on Power and Performance Engineer to help shape power and performance profiling, pre-silicon and post-silicon methodology, TDP definition, workload automation, and product requirement definition for PAI products targeting robotics and automotive markets. This role suits an engineer who enjoys working close to hardware and software, running workloads on Linux-based devices, analysing platform behaviour, building automation harnesses, and using data to influence architecture, product, and platform decisions. We operate with a 10x mindset focused on curiosity, ambition, collaboration, and creating impact. We support hybrid working, with patterns shaped by the teams needs, and we are committed to equal opportunities, mutual respect, and accommodating recruitment needs where possible.

last updated 36 week of 2026