AI Engineer • Computer Vision Engineer • Kernel Developer (C/C++ & Rust)
CUDA Developer for GPU-Accelerated Computing
NVIDIA Inception Member • DeepStream/TensorRT Specialist • YOLO Architect
I'm an AI & Computer Vision Engineer at the 1337 AI Lab (UM6P & OCP Group), where I build real-time GPU-accelerated video analytics systems and low-level deep learning infrastructure.
I work across the full stack of high-performance computing:
- DeepStream 8.0 / TensorRT / CUDA
- GStreamer dynamic pipelines (RTSP/WebRTC)
- YOLOv8, YOLOv11, YOLOv12 (custom implementation)
- Segmentation models (DeepLabv3+, SAM, UNet)
- RabbitMQ for high-frequency metadata
- ONNX export + TensorRT engine optimization
- Low-level systems programming (C/C++ & Rust)
- Kernel development & OS-level GPU interfacing
- Real-time CV pipelines: multi-stream DeepStream with dynamic source add/remove
- YOLO model implementations (stem → C3K2 → R-ELAN → matcher → head → loss)
- High-performance edge + server GPU inference
- Custom CUDA kernels for accelerated computation
- Low-level GStreamer & DeepStream components
- RAG systems with OCR + semantic chunking + embeddings
- Industrial CV solutions: smart parking, helmet detection, conveyor belt inspection, grain analysis, etc.
- 🔥 DeepVisionStream / OcelStream – GPU-accelerated real-time CV platform
- 🧬 YOLOv12 Implementation – full model built from scratch
- 🎥 DeepStream 8.0 IPC – Unix FD-based multi-process inference pipeline
- 📚 RAG System – OCR + semantic chunking + LLM retrieval
- ⚙️ Kernel & Systems Work – experimenting with Rust-based kernel components and GPU-level optimizations
- 📝 Medium: https://medium.com/@ibrahimchifour
- 💼 LinkedIn: https://www.linkedin.com/in/brahim-chifour
Let’s build something crazy with GPUs 🚀