Edge ML Inference Performance Engineer

Wayve; Greater London; Hybrid; GBP 80,000 - 120,000; Full time

Benefits offered by this job

Equity

Relocation support

Hybrid working

L&D budget

Health insurance

Dental

Maternity / Paternity leave

Pension

Job summary

Wayve is building the leading AI platform for autonomous driving with a hybrid work model across London and international offices. The ML performance role focuses on making large transformer models run efficiently on edge devices, balancing latency, memory and power constraints while collaborating with model, platform and deployment teams.

You will optimise inference for edge accelerators like NVIDIA Orin/Thor and Qualcomm platforms, implement kernel-level improvements, and contribute to tooling

Qualifications

• Experience improving performance in production or production-adjacent systems with latency, memory, bandwidth, power, thermal or cost constraints.
• Strong hands-on experience with at least one stack such as TensorRT, CUDA, Qualcomm QNN, Triton or OpenCL.
• Comfortable working from high-level model behaviour down to kernel/runtime-level execution.
• Strong software engineering fundamentals including debugging, profiling, testing and maintainable code.

Responsibilities

• Profiling inference performance across model graphs, compiler/runtime behaviour, kernel execution and memory movement.
• Implementing and validating optimisations in compilers, runtimes and/or kernels, including fusion, scheduling, quantisation-aware performance and custom kernels.
• Building benchmarking and regression tests to track performance across models, devices and software releases.
• Optimising for edge targets such as NVIDIA Orin/Thor and Qualcomm platforms.
• Contributing to team tooling, documentation and technical discussions around ML performance.

Skills

Performance optimization

Python

C++

Cross-team collaboration

Embedded/edge deployment

Tools

TensorRT

CUDA

Qualcomm QNN

Triton

OpenCL

Job description

Wayve is building the leading AI platform for autonomous driving with a hybrid work model across London and international offices. The ML performance role focuses on making large transformer models run efficiently on edge devices, balancing latency, memory and power constraints while collaborating with model, platform and deployment teams.

You will optimise inference for edge accelerators like NVIDIA Orin/Thor and Qualcomm platforms, implement kernel-level improvements, and contribute to tooling