Personal Purpose

This page is created to reflect on my journey through the field of Information Technology and other disciplines such as Machine Learning and Artificial Intelligence. I’ll be continuously uploading my projects that I am working on, solutions of Capture The Flag (CTF) challenges from my own point of view, showcasing my methodology & knowledge.

Technical Skills:

Python | Unix CLI | OST | Bash | SQL | MATLAB | Machine Learning | C (Learning)

Education

M.Sc of Artificial Intelligence | University of East London, London (UEL) (Jan 2024) - (Present)

Major Disciplines:

• Intelligent Systems
• Big Data Analytics
• Machine Learning
• Machine Vision

B.BA of Information Management | Arab Academy for Science, Technology, And Maritime Transport, Egypt (AASTMT) (Sept 2017) - (Sept 2021)

Major Disciplines:

• Information System Analysis & Design
• Website Development
• Database Management
• Networking

Non-Credit Courses & Certificates:

Machine Learning Onramp | Image Processing Onramp | Deep Learning Onramp - MathWorks

Cisco CyberOps (200-201) Assosciate Guide - Udemy

THM Advent of Cyber 2023 - TryHackMe

Capstone Projects

URL Binary Classification (Malicious or Benign) - Supervised ML

Built a model from scratch using pyspark’s framework to determine whether a URL is malicious or benign based on a set of carefully selected features.

Features were identified by the output of a heatmap using numpy’s and sklearn’s libraries that correlated the highest impact of all viable columns to the target label column, and then aggregated by using VectorAssembler to be parsed in as one single feature vector.

Utilized different learners to gain an initial model, furthermore gaining insight to build an approach to tackle the objective, utilized different methods such as aggregating the sum of two different learner outputs, modified losstype etc.

Incorporated ParamGridBuilder and CrossValidator with GBTClassifier, added weight to respective labelled instances as a measure of balance in the dataset.

Finally, a confusion matrix was utilized to measure the accuracy of the trained model’s prediction on the given test set. Achieved 90% accuracy, heavily skewed data

Rice Seed MultiClass-Classification - CNN & Transfer Learning

Built a neural network from scratch as part of my studies for my A.I - Machine Vision course. The network was relatively small in size and performed poorly.

Later on, I decided on using a transfer learning approach to configure a pre-trained neural network that used GoogleNet’s network architechture.

Configured inputLayer and classificationLayer respectively according to the number of classes existing in the dataset, set hyper-parameters to validate the learning process.

Finally, utilized a confusion matrix using MATLAB to analyze the results of true-positives and false-negatives. Achieved 94-95% accuracy

CTF Machine Write-Ups

Basic Pentest Room - THM | Three ST Machine - HTB

Machine Learning Projects

These projects are for self-educational purposes only, some may have very limited application. The following page contains details about each project with links to the rendered code-book and the dataset file that was used in each project.

Supervised / Unsupervised ML Projects

CTF Challenges & Module Walkthroughs

PwnCollege - Modules | OverTheWire - WarGames | CryptoHack - Challenges | PicoCTF - Challenges

Socials & Public Profiles

LinkedIn | TryHackMe | PwnCollege