Staff ML Engineer: Edge AI Tooling & Developer Experience

Arm; Cambridge; Hybrid; GBP 97,000 - 132,000; Full time

Job summary

Arm is seeking a Staff ML Engineer to help build the next generation of ML tooling for developers on Arm architecture. You will lead efforts in model analysis and optimisation, ensuring tools work across ML frameworks and hardware to simplify deployment and execution on Arm-based platforms.

You will design and implement production-grade tooling, collaborate with cross-functional teams, and drive engineering quality with automated tests and iterative feedback from developers.

Qualifications

• Strong understanding of neural-network architecture and model execution.
• Hardware-aware optimization techniques such as quantisation and fusion.
• Experience with PyTorch, ONNX/ORT, TensorFlow or similar runtimes.
• Ability to profile ML workloads and reason about latency/memory trade-offs.
• Passion for building developer-friendly tooling.

Responsibilities

• Join an established team to design and deliver model analysis and optimisation tooling for edge developers via web services and desktop apps.
• Provide technical leadership in ML tooling, shaping architecture and roadmap.
• Work across ML frameworks, runtimes and hardware to deliver intuitive experiences for non-expert developers.
• Collaborate with distributed teams, PMs, and UX to ensure tooling meets real developer needs.
• Drive quality through maintainable software, tests, CI, and customer feedback.

Skills

Neural nets

Model optimization

PyTorch

ONNX

Profiling

Heterogeneous HW

Job description

Arm is seeking a Staff ML Engineer to help build the next generation of ML tooling for developers on Arm architecture. You will lead efforts in model analysis and optimisation, ensuring tools work across ML frameworks and hardware to simplify deployment and execution on Arm-based platforms.

You will design and implement production-grade tooling, collaborate with cross-functional teams, and drive engineering quality with automated tests and iterative feedback from developers.