Mhd Taha Daboul

📚 AI Educational Assistant – RAG System for STEM

Interactive platform using Retrieval-Augmented Generation for Biology, Chemistry, Physics, and Mathematics.

• Integrated subject‑specific textbooks as knowledge bases; rendered equations with LaTeX (KaTeX).
• Designed dual‑mode web interface: “Ask” for answers, “Test” for AI‑generated quizzes.
• Employed embeddings + vector search for accurate context retrieval.

🧠 PharmaScan — Handwritten Prescription Recognition

End‑to‑end pipeline to recognize medicine names from handwritten prescriptions.

• OpenCV preprocessing (grayscale, noise removal, adaptive thresholding, segmentation).
• Combined Tesseract OCR with VGG16 CNN classifier (transfer learning) for higher accuracy.
• Added fuzzy matching for alternative suggestions and drug‑interaction alerts.
• Automated ingestion & retraining workflows with n8n; Docker containerization.

Tech: Python · TensorFlow · OCR · Pandas · n8n · Docker

🔗 GitHub

🏠 House Price Prediction – Regression Models

Exploratory analysis, feature engineering, and model comparison.

• Built and tuned Linear Regression, Random Forest, XGBoost with cross‑validation.
• Improved accuracy via ensemble methods and hyperparameter optimization.
• Delivered lightweight API and business‑friendly Jupyter notebooks.

Tech: Python · Scikit‑learn · XGBoost · Flask · Pandas

🔗 GitHub

🤖 AI Agent Automation (n8n)

Designed intelligent workflows integrating AI models with live APIs.

• Automated structured JSON responses, model retraining triggers, and data pipelines.
• Used Docker for deployment and versioning.