0
Loading Portfolio
Available for innovative roles

Architecting Intelligence from Data.

Currently exploring
Deep Learning
Explore Work ↓ Download CV Contact
Philosophy

Building at the intersection of human cognition and machine intelligence — where data becomes insight and insight becomes action.

George Ehab
George Ehab Sobhy

Computer Science student at E-JUST specializing in Artificial Intelligence and Data Science. Driven by mathematical rigor and practical engineering, I design robust, end-to-end machine learning pipelines — transforming complex, multi-modal data into high-performance predictive systems and scalable deep learning architectures.

7+
ML / DL Projects shipped
3.13
GPA at E-JUST
5+
Domains mastered
26'
Expected grad year
Selected Work

Projects that
push limits.

01
Deep Learning Computer Vision

Brain Tumor Classification

Multiclass MRI classification system using ResNet50 + Transfer Learning. Identifies four tumor types with OpenCV preprocessing and TensorFlow augmentation pipelines.

Complete GitHub
02
Signal Processing Hybrid Architecture

EEG Signal Classification

Hybrid CNN-LSTM model that captures both spatial and temporal features from brain electrical signals. Custom DSP pipeline for cleaning and normalization.

Complete GitHub
03
Machine Learning Feature Engineering

Solar Energy Forecasting

Advanced ML pipeline forecasting solar generation output. Extensive feature engineering with Random Forest and XGBoost algorithm comparison for optimal accuracy.

Complete GitHub
04
Machine Learning Pipeline Architecture

Customer Churn Prediction

End-to-end production ML pipeline using Scikit-learn with ColumnTransformer and RobustScaler. Benchmarked Gradient Boosting against Random Forest with dynamic probability threshold optimization for maximum Recall and F1-score.

Complete GitHub
05
Recommender Systems SVM / Dual Engines

Hybrid Film Recommender Engine

Multi-engine movie recommendation system combining dual similarity modules with an SVM ranking classifier. Features an ensemble hybrid engine that balances cosine content matching and collaborative feature weighting.

Complete GitHub
06
Image Restoration SRResNet

Image Super-Resolution

DL model enhancing image resolution with SRResNet architecture. Evaluation via PSNR and SSIM metrics. Actively improving restoration fidelity.

In Progress
07
NLP ASR

Arabic ASR System

Research-driven Arabic Automatic Speech Recognition system targeting improved Word Error Rate via noise injection and beam search optimization techniques.

In Progress
Arsenal

Tools that
build futures.

Core ML & Data
Python Pandas NumPy Scikit-learn SQL Regression Classification Clustering XGBoost Random Forest
AI.
Built at the edge of
machine cognition.
Deep Learning
TensorFlow Keras CNNs RNNs / LSTMs ResNet50 Transfer Learning SRResNet
Advanced Domains
Reinforcement Learning Q-Learning EEG / Signal DSP NLP / ASR Image Restoration
Tools
Git & GitHub VS Code OpenCV
Mathematics
Statistics for ML Linear Algebra Calculus
Career & Education

Milestones
that shaped me.

2025
Mindset
AI Diploma — Completed
Comprehensive advanced AI program covering modern deep learning architectures, NLP, computer vision, and production ML workflows.
2025
Research Projects
Independent AI Researcher
Arabic ASR system development and SRResNet-based image restoration — both active research initiatives exploring frontier ML challenges.
2022
E-JUST University
B.Sc. Computer Science, AI Specialization
Egypt-Japan University of Science and Technology, Alexandria. GPA 3.13. Specializing in artificial intelligence, data science, and intelligent systems design. Expected graduation 2026.
Pre-2022
IEEE · ISEF · ISEAC
Scientific Research Competitor
Represented at international science competitions during high school — building a foundation in systematic research, innovation, and engineering thinking.
Email copied!