CareerPlan

ML Runtime Engineer

Toulouse, OccitanieFull-time2026-07-01 → 2026-11-19
Based in France

Core

Compile ONNX models into optimized engines for satellite accelerators and build performance tooling to validate models against power, thermal, and latency budgets for flight hardware.

Role type

ML Runtime Engineer (Embedded/Edge)

Builds

Optimized inference engines for accelerators (TensorRT, Hailo, AMD/ROCm), Space Inference Engine execution providers, and SDKs for model packaging.

Domain

Aerospace / Satellite AI / Embedded Systems

Deliverable

production ML models

Required skills

C++, Python, Model compilation (TensorRT, ONNX Runtime), Quantization, Embedded GPU deployment (NVIDIA Jetson), Benchmarking & profiling, CI/CD with hardware-in-the-loop

Preferred skills

Hailo SDK, AMD ROCm, Custom kernel writing, RF/IQ signal data, Meson build, Yocto

Technologies

ONNX, TensorRT, Hailo, AMD ROCm, NVIDIA Jetson, Meson, Yocto

Responsibilities

Own model compilation for various accelerators within strict power/thermal budgets; Build ground-side compile and delivery paths; Develop performance tooling for benchmarking and profiling; Contribute to Space Inference Engine and SDK; Coordinate with OBSW/Runtime teams.

Seniority

Mid-Senior, hands-on IC