Ingénieur Performance IA – C++/CUDA (H/F)
Core
Develop and optimize real-time AI perception algorithms on embedded GPUs for autonomous robots and advanced targeting systems.
Role type
Senior IC embedded AI performance engineer (C++/CUDA)
Builds
High-performance AI inference modules for defense, aerospace, and industrial robotics
Domain
Defense & Aerospace / Embedded AI / GPU Computing
Deliverable
production ML models
Required skills
C++17/20, CUDA, embedded GPU architecture, real-time systems, computer vision models, model optimization (quantization, pruning), profiling tools (Nsight, CUDA Profiler)
Preferred skills
Aerospace/defense/robotics industry experience, operational deployment on embedded GPUs, thermal/power constraint management, open source contributions
Technologies
C++17/20, CUDA, TensorRT, DeepStream, Nsight Systems, Nsight Compute
Responsibilities
Design and optimize AI/ML modules for embedded constraints; Profile and resolve GPU bottlenecks; Tune CUDA kernels and adapt models for performance; Validate real-time latency and stability; Integrate solutions with heterogeneous systems; Document and train teams on GPU optimization best practices
Seniority
Senior, hands-on IC
