Postdoctoral Researcher at ONERA (The French Aerospace Lab), Toulouse, France.
My research focuses on embedded artificial intelligence, robust deep learning, and AI-hardware co-design for constrained and safety-critical systems. I currently work on the design and certifiable FPGA deployment of lightweight neural networks on programmable logic.
I received my Ph.D. in Computer Engineering from Polytechnique Montréal, where I worked on neural-network-based spacecraft pose estimation, from single-image estimation to temporally consistent video processing, as well as real-time deployment on CPU, GPU, and FPGA platforms.
- Embedded AI
- AI-hardware co-design
- Robust and reliable deep learning
- FPGA acceleration
- Real-time inference on constrained platforms
- Visual perception for aerospace and intelligent mobility
- Mobile-URSONet — Lightweight neural network for onboard spacecraft pose estimation
- FPGA-SpacePose — FPGA implementation and mixed-precision quantization for real-time inference
- Spacecraft-Pose-Estimation-Framework — Training, deployment, and dataset generation tools
julien.posso [at] gmail [dot] com