Building machines that see — from research to production
End-to-end computer vision and deep learning systems — 18 production-grade projects spanning medical imaging, 3D reconstruction, multimodal AI, generative vision, and autonomous perception.
| Domain | Project | Key Metric |
|---|---|---|
| 🏥 Medical Imaging | MediScan — DenseNet-121 chest X-ray | 97.3% acc · AUC 0.99 |
| 🔍 Object Detection | UrbanSense — YOLOv8 + INT8 ONNX | 45 FPS · mAP 0.891 |
| ⚡ ML Serving | InferEdge — Triton + ONNX server | <8ms P99 · 500+ req/min |
| 🌐 3D Reconstruction | NeRFusion — 3D Gaussian Splatting | PSNR 34.1 dB · 120 FPS |
| 🎬 Video Segmentation | SAMFlow — SAM-2 + optical flow | J&F 87.3% · zero-shot |
| 🧠 Multimodal AI | LangVision — LLaVA-1.6 VQA | VQAv2 82.1% · 35 tok/s |
| 🎨 Generative AI | DiffCtrl — ControlNet + SDXL | FID 8.2 · 94% user pref |
| ✂️ Segmentation | SegForge — U-Net + ASPP | mIoU 0.843 |
| 📄 OCR | DocuAI — CRNN + CRAFT Arabic/Latin | 96.4% word accuracy |
| 🔎 Visual Search | VisionLang — CLIP + FAISS | Recall@5 92.4% |
| 👤 Face Recognition | FaceVault — ArcFace | TAR 99.2% @ FAR 0.1% |
| 📹 Tracking | TrackMaster — DeepSORT | MOTA 0.812 |
| 🕺 Pose Estimation | PoseForge — ViTPose | AP 77.1% |
| 🎭 GAN Generation | SynthVision — StyleGAN2 | FID 18.4 |
| 🔭 Super-Resolution | SuperRes — ESRGAN | PSNR 32.7 dB |
| 🎯 Action Recognition | ActionNet — SlowFast | Top-1 87.4% |
| 📏 Depth Estimation | DepthPro — DPT | δ1 93.2% |
| 🚗 Autonomous Perception | AutoPerception — PointPillars BEV | mAP 0.624 nuScenes |
Core: PyTorch · TorchVision · ONNX Runtime · OpenCV · NumPy · CUDA
CV & DL: YOLOv8 · DenseNet · ViT · CLIP · LLaVA · CRNN · U-Net · ArcFace · SlowFast · DPT · PointPillars · 3D-GS · SAM-2 · SDXL · ControlNet
Serving: FastAPI · Triton Inference Server · vLLM · Docker · INT8 Quantization · TensorRT · xFormers
Languages: Python · C++ (OpenCV/CUDA) · SQL · Bash
- Freelance / Contract — Fixed-price or hourly, starting $45/hr
- Remote Full-time — AI Engineer / ML Engineer / Computer Vision Engineer
- Consulting — CV pipeline architecture, model optimization, inference serving
📩 sherifabdelrady@gmail.com · Portfolio · LinkedIn
Cairo, Egypt · Responding within 24 hours · Available worldwide · UTC+2