Staff ML Performance Engineer: Edge Inference & Systems

AI Startups UK; Greater London; Hybrid; GBP 120,000 - 170,000; Full time

Job summary

Wayve is hiring a Staff ML Performance Engineer to optimise ML inference on edge devices, enabling large transformer models to run efficiently on cost-effective hardware for our autonomous driving product.

The role spans ML systems, compilers, runtimes, and embedded deployment, with hands-on work across multiple platforms and close collaboration with model developers to shape architecture and tooling.

Qualifications

• Proven ability to optimise production systems under tight constraints (latency, memory, power).
• Familiar with ML stacks and rapid learning of new frameworks.
• Able to operate across abstraction levels from model behavior to kernel execution.
• Strong software engineering fundamentals (debugging, profiling, testing).
• Excellent communication and cross-team collaboration.

Responsibilities

• Identify, implement and validate optimisations in ML compilers, runtimes, and kernels (e.g. operator fusion, scheduling, quantisation).
• Profile and pinpoint bottlenecks across the full inference stack and deliver measurable improvements.
• Build benchmarking and regression tests to ensure performance across models and devices.
• Develop and optimise for multiple target platforms (e.g. NVIDIA Orin/Thor, Qualcomm).
• Collaborate with model developers to influence architecture and deployment decisions.
• Contribute to roadmaps and tooling and raise performance standards across the team.

Skills

Performance optimization

CUDA/TensorRT

Python

C++

Profiling

Communication

Tools

TensorRT

CUDA

Qualcomm QNN

Triton

ONNX

Job description

Wayve is hiring a Staff ML Performance Engineer to optimise ML inference on edge devices, enabling large transformer models to run efficiently on cost-effective hardware for our autonomous driving product.

The role spans ML systems, compilers, runtimes, and embedded deployment, with hands-on work across multiple platforms and close collaboration with model developers to shape architecture and tooling.