Dana Alrijjal

Dana Alrijjal

About Me

Hello! I'm Dana, an expected Computer Science graduate from Effat University, specializing in Artificial Intelligence. Over the years, I’ve deepened my expertise in AI, programming, and problem-solving, and I’m passionate about applying technology to real-world challenges.

Throughout my academic journey, I’ve led and contributed to diverse projects, from network design to intelligent systems and data-driven applications. I also engage in projects that combine AI with cybersecurity principles to build secure and adaptive solutions. My strong foundation in mathematics, combined with a hands-on approach to learning, helps me tackle complex problems with creativity and precision.

As I approach graduation, I’m actively exploring opportunities to grow professionally and contribute to impactful technological innovations.

Get in Touch

Education

Bachelor's Degree in Computer Science

Effat University

August 2022 - Present

GPA: 3.97

50% Merit Scholarship

Dean's List — 2022 to 2026

American High School Diploma

Coral International School

September 2008 - May 2022

GPA: 4.0

Certifications & Training

Anthropic Courses

Complete

Claude 101

Learn how to use Claude for everyday work tasks, understand core features, and explore resources for more advanced learning.

Complete

Claude Code 101

Learn how to use Claude Code effectively in your daily development workflow.

Complete

Model Context Protocol: Advanced Topics

Advanced MCP implementation patterns including sampling, notifications, file system access, and transport mechanisms for production MCP server development.

Complete

AI Fluency: Framework & Foundations

Learn to collaborate with AI systems effectively, efficiently, ethically, and safely.

Experience

Important Courses

Principles of Computing

Fall 2022

Database Systems

Spring 2023

Advanced Programming

Spring 2023

Object Oriented Programming

Fall 2023

Web Development

Fall 2023

Computer Networks

Spring 2024

Operating Systems

Spring 2024

Data Structure

Spring 2024

Artificial Intelligence

Fall 2024

Data Sciences

Fall 2024

Software Engineering

Spring 2025

Machine Learning

Spring 2025

Natural Langauge Processing

Fall 2025

Advanced Topics in AI

Fall 2025

Big Data Analytics

Spring 2026

Algorithm Analysis

Spring 2026

Projects

Adversarial Robustness in Cyber Threat Intelligence (CTI)

An applied AI security project focused on improving the robustness of NLP models used for cyber threat intelligence classification.

  • Fine-tuned a DistilBERT model for CTI report classification
  • Implemented FGSM adversarial training to defend against textual perturbation attacks
  • Evaluated baseline vs adversarially trained models using Clean Accuracy and Macro F1
  • Measured robustness using Robust Accuracy and Attack Success Rate (ASR)
  • Applied adaptive defense concepts to enhance model reliability in security-critical environments

The project demonstrates how adversarial training can significantly improve the resilience of NLP models deployed in real-world cybersecurity pipelines.

View on GitHub

Adversarial-Resilient Big Data Pipeline for CTI

A distributed pipeline for processing Cyber Threat Intelligence data using Apache Spark.

  • Built a 7-stage pipeline (ETL, anomaly detection, adaptation, ML, and audit)
  • Processed 11K+ records using Spark and Parquet
  • Implemented anomaly detection and data cleaning, reducing noise by 50%
  • Trained a TF-IDF + Logistic Regression model (~82% accuracy)
  • Evaluated robustness under adversarial conditions

This project extends adversarial robustness from model-level to pipeline-level, improving reliability in real-world cybersecurity data systems.

View Project

SecureTwin Cybersecurity Digital Twin

A proactive cybersecurity testing platform that enables enterprises to optimize security defenses through:

  • Real-time attack simulation with 5+ predefined attack types
  • AI-driven threat intelligence and dynamic defense adjustments
  • Digital twin network modeling for risk-free testing
  • Compliance tracking with GDPR and ISO 27001 standards
  • Role-based access control for secure system management

The application achieved 10,000 node scalability with 100ms latency for real-time responses during testing, demonstrating robust performance for enterprise security needs.

View on GitHub

IoT Cyber Attack Detection Using ML

Machine learning-based Intrusion Detection System (IDS) for IoT networks featuring:

  • Data preprocessing (missing value imputation, normalization)
  • Dimensionality reduction with PCA
  • Class balancing using SMOTE
  • Multiple model training (Random Forest, SVM, XGBoost)
  • 95% accuracy in attack detection

The system demonstrates how ML can provide scalable security solutions for vulnerable IoT ecosystems.

View on GitHub

Predicting Student Performance with ML

Data science project aimed at predicting secondary school students' final grades (G3) using demographic, academic, and behavioral factors.

  • Dataset of 1,044 students from two Portuguese schools
  • Preprocessing: Label encoding, feature scaling, and train-test split
  • Exploratory Data Analysis: Correlation heatmaps, feature distribution
  • Model training: Linear Regression and Random Forest Regressor
  • Evaluation metrics: MAE, RMSE, and R² Score

Results showed that prior grades (G1, G2) were the most significant predictors of final performance. The project offers insights to support educational policy and early intervention strategies.

View on GitHub

Virtual Art Gallery Website

Museo is a dynamic virtual art gallery designed to provide an engaging online experience for art enthusiasts. Key features include:

  • Responsive design with navigation bar and drop-down menu
  • Six distinct art categories showcasing renowned pieces
  • Detailed artist pages and interactive search functionality
  • Built with HTML, CSS, JavaScript, Bootstrap and Font Awesome
  • Full CRUD system with secure admin dashboard

This project enhanced our skills in creating interactive, database-driven websites despite challenges in database connectivity and responsive design.

View on GitHub

Network Design for Educational Institution

A comprehensive network plan for a medium-sized educational institution with:

  • Three distinct campuses (North, South, Central) each with dedicated VLANs
  • Centralized switches, routers and access points
  • Administrative hub with inter-VLAN communication
  • DHCP and DNS services integration
  • VLAN segmentation for enhanced security

The design optimizes technological infrastructure to support academic and administrative functions through secure, scalable networking.

View on GitHub

To-Do-List Application

A digital solution for task management and time organization featuring:

  • Minimalist and user-friendly interface
  • Support for simple, deadline-driven, and recurring tasks
  • Task organization into customizable categories
  • Mark tasks as completed functionality
  • Edit or delete tasks and categories

The application focuses on streamlined task management to enhance productivity and organizational efficiency.

View on GitHub

Papers

Advancing Cybersecurity with Digital Twin Technology

This research explores the transformative potential of Digital Twin (DT) technology in cybersecurity, proposing an innovative framework that integrates IoT sensors, AI, and machine learning for proactive threat detection. Co-authored by Dana Alrijjal and Jouri Aldaghma, the study demonstrates a digital twin model capable of simulating cyber-attacks (ransomware, DDoS, phishing) with 97.5% detection accuracy and sub-1.5 second latency. The framework significantly outperforms traditional security methods, reducing false positives by 55% while maintaining low computational overhead. The research addresses critical challenges in data synchronization and scalability, validated through both simulated environments and real-world smart grid deployment. This work contributes to Saudi Vision 2030's digital transformation goals by enhancing infrastructure resilience against evolving cyber threats.

View on ResearchGate

A Multi-Layered Adaptive Framework for Adversarially Robust AI in Cybersecurity

This paper presents a comprehensive study on adversarial machine learning threats targeting AI systems in cybersecurity, with a particular focus on NLP-based Cyber Threat Intelligence (CTI). Co-authored by Dana Alrijjal and Jouri Al Daghma under the supervision of Dr. Naila Marir, the research analyzes state-of-the-art adversarial attacks and defenses published between 2022–2025.

The study proposes a Multi-Layered Adaptive Defense Framework that combines adversarial detection, incremental retraining, and explainable auditing into a closed feedback loop. The framework is validated through a practical case study using a DistilBERT classifier trained on CTI reports, where FGSM adversarial training significantly improves robustness while maintaining competitive clean accuracy. This work contributes to the development of trustworthy and resilient AI systems for security-critical applications.

View on ResearchGate

Securing the Internet of Things: Addressing Cross-Layer Vulnerabilities

This study explores security and privacy challenges in IoT ecosystems through a systematic examination of the layered IoT architecture. Co-authored by Dana Alrijjal and Jouri Aldaghma, the research evaluates vulnerabilities and solutions across Perception, Network, Middleware, and Application layers. Key findings reveal critical issues including weak authentication mechanisms, DoS attacks, and cryptographic inefficiencies, alongside promising solutions like lightweight encryption and blockchain integration. The paper highlights the fragmented nature of current security approaches and emphasizes the need for integrated frameworks to address cross-layer vulnerabilities. This work contributes to global IoT security efforts in healthcare, smart cities, and industrial IoT by proposing scalable, robust security strategies that enhance system reliability and trust while accommodating resource constraints.

View on ResearchGate

Wearables and Augmented Humans: Ethical and Regulatory Implications

This paper examines the ethical and regulatory challenges of wearable technologies and human augmentation devices. Co-authored by Dana Alrijjal and Jouri Aldaghma, the research explores how these technologies enhance physical, cognitive, and sensory capabilities while raising significant concerns about privacy, autonomy, social equity, and psychological well-being. Through a systematic review of ten peer-reviewed papers, we analyze these issues using ethical frameworks like Utilitarianism and Kantianism. The study also evaluates regulatory approaches in Saudi Arabia (PDPL) and internationally (GDPR, HIPAA), identifying gaps in privacy protections and equitable access. Key recommendations include strengthening data privacy laws, promoting ethical design practices, and developing harmonized global standards to ensure responsible adoption of these transformative technologies.

View on ResearchGate

AI Driven Virtual Environments

This research paper explores the intersection of Artificial Intelligence (AI) and virtual environments. Co-authored by Celine Al Harake and Dana Al Rijjal under the supervision of Dr. Khadija Itani, the study examines the potential and challenges of integrating AI into virtual spaces. The paper delves into how AI can enhance virtual environments across various fields like education, healthcare, and social interaction, while also addressing ethical concerns and the impact on human well-being. Through questionnaires and interviews, the research highlights both the transformative possibilities and the anxieties associated with these technologies, advocating for responsible development and clear guidelines to ensure ethical and beneficial use of AI in virtual settings.

View on ResearchGate

Security Mechanisms in Modern Software Systems

This systematic review analyzes modern software security mechanisms through the lens of sustainable development, evaluating 20 peer-reviewed studies (2022-2024) using PRISMA methodology. Co-authored by Dana Alrijjal and Jouri Aldaghma, the research demonstrates how lattice-based post-quantum cryptography (CRYSTALS-Kyber) achieves 1,800 ops/sec at 192-bit security while reducing energy consumption by 37%, directly supporting SDG 9's infrastructure targets. The study reveals Zero Trust implementations reduce successful cyberattacks by 68-72% in government systems (advancing SDG 16.6), while memory-safe languages decrease vulnerability density by 89% in critical infrastructure (supporting SDG 16.4). The paper identifies significant gaps in developing-country implementations and proposes a novel framework for SDG-aligned security assessments, providing both technical insights and policy recommendations for sustainable cybersecurity.

View on ResearchGate

Advancements in Kernel Concurrency: Leveraging Machine Learning for OS Innovation

Co-authored by Dana Alrijjal and Jouri AlDaghma, this paper presents a comprehensive review of cutting-edge research in kernel concurrency, emphasizing the integration of machine learning techniques to address longstanding challenges in operating system (OS) development. The study explores novel frameworks and methodologies introduced in recent research papers that utilize machine learning to enhance kernel concurrency control, testing strategies, and bug detection. Key contributions include SynCord, a framework for application-informed kernel synchronization, and Snowboard, a tool for systematic inter-thread communication analysis to uncover concurrency bugs. This review highlights the transformative potential of machine learning in optimizing performance and security within OS kernels, while also addressing future research directions such as customizable kernel mechanisms, security enhancements, and scalability concerns.

View on ResearchGate

An Adaptive Adversarial-Resilient Big Data Pipeline for Cyber Threat Intelligence Classification

This paper presents A3L-BD, a seven-stage Spark-based pipeline for adversarially resilient Cyber Threat Intelligence (CTI) processing. Co-authored by Dana Alrijjal, Jouri Aldaghma, and Dr. Naila Marir, the work formalises a black-box data poisoning threat model and operationalises it through an adaptive five-flag anomaly detection algorithm. Evaluated on the AnnoCTR corpus (11,114 records annotated with MITRE ATT&CK labels), the pipeline reduces training noise by 51.7% and demonstrates measurable classification recovery under adversarial conditions (82.25% clean accuracy; 81.81% under attack; 81.86% after defence), completing a full pipeline run in 126.9 seconds with reproducible audit artefacts. This work extends adversarial robustness principles from the model layer to the distributed data engineering layer of CTI pipelines.

View on ResearchGate

Skills

Core Skills

Problem Solving
Data Analysis
Project Management
Critical Thinking
Time Management
Communication
Teamwork
Research
Mathematics
Statistics
Data Visualization
System Design
Algorithm Design

Technologies

Python
Java
C++
HTML/CSS
JavaScript
SQL
Git
Pandas
NumPy
TensorFlow
Matplotlib
Cisco Packet Tracer
Jupyter Notebook
Scikit-learn

Contact