Portfolio
Paul
FOLTZER
Engineering student at Mines Saint-Étienne (ISMIN) — Embedded Systems & Security of Digital and AI Architectures
I build systems where electronics, mechanics and software meet — from embedded firmware to data processing.
5documented technical projects
9 000+test assertions written (TransformerCpp)
×30training speed-up after the GPU port
1research paper under review
Selected work
Featured projects
Toolbox
Technical skills
Deep learning
- PyTorch / MONAI
- 3D volumetric segmentation (Swin-UNETR)
- Transfer learning & self-supervised pre-training
- Class imbalance (weighted and focal losses)
AI & high-performance computing
- C++17 / CUDA
- From-scratch implementations (Transformer, CNN)
- cuBLAS, custom GPU kernels
Embedded
- STM32 / C
- Real-time firmware
- Stepper motor control (DRV8825, NEMA17)
Systems & Linux
- Jetson / L4T
- Device tree & overlays
- Low-level hardware debugging
Data & imaging
- Python (numpy, scipy, scikit-image)
- Computed tomography imaging
- Tomographic reconstruction (iradon)
- Statistical analysis (Wilcoxon, Wasserstein)
Fabrication & instrumentation
- Parametric CAD in Python (trimesh)
- FDM 3D printing
- X-ray CT acquisition
Tools
- Git
- STM32CubeIDE
- 3D Slicer
- VGStudio
- VS Code
Let's talk about your project
Internship, work placement or plain technical curiosity — my inbox is open.