Building intelligent, data-driven systems with a strong focus on real-world model deployment and end-to-end AI pipelines.
I am a results-driven AI Engineer in the making, specializing in Machine Learning, Deep Learning, and Computer Vision. I bridge the gap between raw data and intelligent systems — building models that classify, detect, predict, and generate, with a strong focus on real-world deployment and hands-on impact.
Philosophy: "Don't just train the model — understand it, deploy it, and make it matter."
My AI journey spans the full spectrum — from classical Machine Learning (supervised and unsupervised algorithms including SVM, XGBoost, Random Forest, and more) to Deep Learning with neural networks, and Computer Vision using CNNs, transfer learning, YOLO, and OpenCV. I have applied these skills across multiple hands-on projects including real-time image classification systems, object detection pipelines, and predictive analytics models.
| Project | Description & Key Features | Technical Highlights | Links |
|---|---|---|---|
| 🔵 Akina-Draw | Real-time hand-drawing classifier — predicts what you draw while the pen is still moving. Live confidence score updated per stroke. Interactive desktop UI with 89 curated classes, brush/eraser controls, undo, and class reference panel. Showcased at RoboTech ASU Fair '26 & AZEX Competition (May 2026) | Inception CNN, Google Quick Draw Dataset (89/300+ classes), OpenCV canvas preprocessing, Pygame UI |  |
| 🟣 SpectraSense | Intelligent conveyor-belt sorting system classifying objects as metal or non-metal across 4 streams: Metal Electronics, Metal Accessories, Metal Tools, Non-Metal. Hardware-integrated with servo motors and sensors for automated physical sorting. Showcased at RoboTech ASU Fair '26 & AZEX Competition (May 2026) | YOLO, OpenCV, Roboflow dataset, ESP32-CAM + ESP32 + Arduino Uno, MG996R Servo Motors, Hall Effect Sensors | |
| 🟢 Bank Churn Predictor | End-to-end churn prediction pipeline on banking customer data. Identifies customers at risk of leaving using classical ML and Deep Learning. Full feature engineering, class imbalance handling, and comparative model evaluation with ROC-AUC reporting. | scikit-learn, Deep Learning, pandas, Feature Engineering, F1-Score & ROC-AUC Evaluation | |
| 🟡 Telco Churn Predictor | Churn prediction system on telecom data. Deep EDA uncovering patterns across contract types, tenure, and service usage. Ensemble methods + neural network with hyperparameter tuning, cross-validation, and feature importance analysis. | scikit-learn, Deep Learning, EDA, Hyperparameter Tuning, Cross-Validation, pandas | |
| 🔴 Student Mental Health Predictor | Classifies student mental health risk levels from academic and behavioral survey data. Includes a fully interactive web-based GUI built with NiceGUI for real-time end-user prediction without any code. Showcased at RoboTech ASU Fair '26 | scikit-learn, NiceGUI, Classification, Feature Engineering, pandas | |
- Best Member of November — RoboTech ASU
- Best Machine Learning Task — RoboTech ASU
- Best Deep Learning Task — RoboTech ASU
- Getting Started with Deep Learning — NVIDIA
- AI & ML Bootcamp — GDG on Campus Hurghada
- Machine Learning Summer Training — NTI
- AI Fundamentals — DataCamp
- Participating with SpectraSense — AZEX (May 2026)
- Participating with Akina-Draw — AZEX (May 2026)
