Pixel Cat Yarn Ball

Nadine Khattab

I'm a Computer Science student at AAST focused on finding patterns in complex data. I build models that predict, classify, and recommend, turning raw data into actionable insights. To make sure my models actually work in the real world, I use my AWS Cloud experience to build serverless pipelines that scale effortlessly.

Data Science Machine Learning Predictive Modeling AWS Cloud Engineering

Education

B.Sc. in Computer Science

2026

Arab Academy For Science, Technology, and Maritime Transport

Relevant Coursework

Algorithms Database Systems Networks Operating Systems Web Dev

Tech Skills

Languages

Python, Java, Kotlin, C, SQL

ML / AI

TensorFlow, Scikit-learn, RAG

Cloud

AWS Lambda, S3, DynamoDB, API Gateway, EventBridge, CloudWatch

Automation / Scraping

Selenium, Requests, BeautifulSoup

Data / Libraries

Pandas, NumPy, Matplotlib

Dev Tools

Linux, Git, VS Code, Jupyter

MemoriaHome

MemoriaHome Project Preview
Machine Learning Face Recognition Vocal Distress Recognition Stutter Recognition

Tech Used

InsightFace Yolov8n SpeechBrain Wav2vec Cloudflare R2 Kinect SDK NestJS MQTT Android SensorManager Google Health Services

MemoriaHome is a real-time dementia patient monitoring system designed to ease the burden on caregivers in both private homes and nursing facilities. Using a Microsoft Kinect v2 for visual and acoustic input and a Samsung Watch 7 for physiological data, the system runs InsightFace for facial recognition and tracking, Kinect native skeletal tracking & YOLOv8n for fall detection, and SpeechBrain combined with Wav2vec to detect vocal distress markers such as calls for help, elevated pitch, or stuttering. Caregivers receive real-time alerts through a unified dashboard, making it easier to stay aware of the patient's condition without constant supervision

View on Github →

Latest Projects

AI RESUME RANKER
AWS Lambda NLP

Architected a serverless NLP pipeline using Amazon Comprehend to automate resume keyword analysis and ranking.

AWS Cloud AI Project
RAG PIPELINE
ChromaDB Vector Search

Built a Retrieval-Augmented Generation system using vector embeddings for semantic, context-aware AI interactions.

View on Github →
Walmart Sales Forecast Project Preview
LightGBM Time Series

Engineered time-series lag and rolling average features to forecast weekly sales with high precision.

View on Github →
Traffic Vision Project Preview
CNN Computer Vision

Trained a CNN model on the GTSRB dataset with data augmentation for robust real-world image classification.

View on Github →
MOVIE RECOMMENDATION SYSTEM
Collaborative Filtering

Developed user-based and item-based collaborative filtering engines to recommend content via rating behaviors.

View on Github →
LOAN APPROVAL PREDICTOR Project Preview
SMOTE Binary Classification

Handled class imbalance with SMOTE to train high-accuracy predictive models for financial loan risk assessment.

View on Github →