In-Space ML Runtime Engineer

Academic Key; Toulouse; Híbrido; EUR 70.000 - 120.000; Jornada completa

Ventajas ofrecidas por este puesto de trabajo

Equity

Paid time off

Health and life insurance

Lunch vouchers

Travel opportunities

Offsites

Relocation assistance

Relocation assistance to Toulouse

Descripción de la vacante

Loftorbital is seeking a ML Runtime Engineer in Toulouse to own model compilation and performance tooling for running ML models on flight hardware. You will optimize ONNX models for accelerators like TensorRT and Jetson, while ensuring power, thermal, and latency budgets are met.

You will work across model compilation, runtime internals, quantization, and hardware‑in‑the‑loop CI, often in the same day, collaborating with the Space Inference Engine and SDK teams to push the envelope of onboard AI

Formación

• Strong experience in systems and ML deployment.
• Experience turning ONNX models into optimized engines for edge devices.
• Ability to work across hardware-in-the-loop tests and CI/CD pipelines.

Responsabilidades

• Own model compilation: turn partner ONNX models into optimized engines per accelerator within power & thermal budgets.
• Build the ground-side compile & delivery path; define on-node mismatch detection/rejection.
• Build performance tooling: benchmarking, profiling, operator-coverage matrices, budget validation.
• Contribute to Space Inference Engine and the SDK; coordinate OBSW / Runtime team.

Conocimientos

C++ & Python

Model compilation

Embedded GPU

Profiling & tooling

Meson Yocto

Herramientas

TensorRT

ONNX Runtime

Jetson

Yocto

Meson

Descripción del empleo

Loftorbital is seeking a ML Runtime Engineer in Toulouse to own model compilation and performance tooling for running ML models on flight hardware. You will optimize ONNX models for accelerators like TensorRT and Jetson, while ensuring power, thermal, and latency budgets are met.

You will work across model compilation, runtime internals, quantization, and hardware‑in‑the‑loop CI, often in the same day, collaborating with the Space Inference Engine and SDK teams to push the envelope of onboard AI