AI/ML Engineer · Deep Learning · NLP · Full-Stack Developer
NLP Text Classification System
Multi-class text classification with BERT + LSTM, deployed as a real-time Flask API.
Professional Experience
IT Support Specialist · Al-Muna Kareem Hotel
Dec 2025 – PresentDeliver day-to-day IT support for hotel operations, resolving hardware, software, network, and user issues. Maintain IT equipment and systems, assist employees with technical requests, and ensure reliable operation of essential business services.
Building Intelligent Systems, Solving Real Problems
I'm Mohammed Mahrous — an AI/ML Engineer who turns data into intelligent, production-ready systems. I architect end-to-end machine learning pipelines, design and train deep learning models for NLP, Computer Vision, and Predictive Analytics, and deploy them at scale using frameworks like TensorFlow, PyTorch, and Scikit-learn.
I also build the web applications that bring these models to life — full-stack development from responsive frontends to robust backend APIs. My IT infrastructure background means I understand what it takes to keep systems secure, optimized, and running 24/7.
Verified Credentials
Formal training backing up the hands-on project work.
IBM AI Engineering Professional Certificate
IBM
Fundamentals of Artificial Intelligence
Saudi Data & AI Authority (SDAIA)
Introduction to Data Science
Cisco Networking Academy
Deep Dives
Problem, approach, and results for the projects that best show how I work.
NLP Text Classification System
Multi-class text classification with BERT + LSTM, deployed as a real-time Flask API.
Read case study →Customer Churn Prediction
End-to-end churn prediction pipeline with ensemble models and class-imbalance handling.
Read case study →Image Classification with CNN
Custom CNN architecture for image classification, trained with data augmentation.
Read case study →Sentiment Analysis System
Sentiment classifier for customer reviews, benchmarking three supervised models.
Read case study →Recommendation Engine
Collaborative filtering recommender combining content-based and hybrid approaches.
Read case study →Credit Card Fraud Detection
Fraud detection on a highly imbalanced dataset, benchmarking an SVM against a Decision Tree with ROC-AUC.
Read case study →Additional Work
IBM ML Certificate — Lab Work
11 completed, annotated, and fully executed labs from the IBM Machine Learning with Python certificate: KNN and classification metrics, Random Forest vs XGBoost, Ridge/Lasso regularization, K-Means / DBSCAN / HDBSCAN clustering, PCA / t-SNE / UMAP, and Scikit-learn pipelines with GridSearchCV. Organized by topic, with attribution to the original IBM materials.
Drug Classification with Decision Trees
Multi-class Decision Tree classifier that recommends one of five drugs from patient features (age, sex, blood pressure, cholesterol, Na-to-K ratio). Label-encoded categorical features, analyzed feature correlation and class balance, trained and visualized the tree, and evaluated accuracy on a held-out test set.
NYC Taxi Tip Regression Tree
Decision Tree Regressor predicting taxi tip amount from NYC yellow-cab trip features. Ran correlation analysis to drop uninformative columns, L1-normalized the feature matrix, evaluated with MSE and R-squared, then compared tree depths to study the bias-variance tradeoff.
House Price Prediction
Implemented a multiple linear regression model for housing price prediction using scikit-learn. Performed extensive EDA, feature engineering, multicollinearity analysis, and model evaluation with MAE, RMSE, and R-squared metrics.
Spam Detection System
Built a text classification model to detect spam messages using supervised machine learning and NLP feature engineering. Implemented Bag of Words, TF-IDF vectorization, and evaluated with precision, recall, and F1-score.
Simple Linear Regression
Predictive modeling using Simple Linear Regression. Analyzed relationships between variables, visualized data trends, and evaluated accuracy with MAE, MSE, and R-squared metrics on real-world datasets.
Logistic Regression Classifier
Binary classification model using Logistic Regression with data preprocessing, feature scaling, and model training. Evaluated with confusion matrix, precision, recall, F1-score, and ROC-AUC curve analysis.
Data Visualization Dashboard
Interactive data analysis dashboard with dynamic charts, filters, and statistical summaries. Built with Python visualization libraries for exploratory data analysis on real-world datasets.
Full-Stack Portfolio Website
Modern responsive portfolio website with interactive particle animations, dynamic GitHub integration, and smooth scroll-triggered transitions. Built with vanilla JS, CSS animations, and GitHub Pages deployment.
IT Network Monitoring Tool
Automated network monitoring script tracking uptime, latency, and bandwidth across infrastructure. Real-time alerts and logging for system administration and troubleshooting.
Latest Projects
Automatically synchronized with my GitHub repositories.
ibm-ml-course-labs
Completed & annotated labs from the IBM Machine Learning with Python certificate: classification, tree ensembles, clustering, PCA/t-SNE/UMAP, pipelines & tuning.
Created Aug 2026 · Updated Aug 2026
View Repository →machine-learning-projects
Three end-to-end ML projects: customer churn (XGBoost/SMOTE), image classification (CNN), and a movie recommender (SVD). Fully executed notebooks.
Created Jul 2026 · Updated Jul 2026
View Repository →nlp-sentiment-classification
NLP sentiment classification: benchmark of text representations (TF-IDF -> BERT) + a Flask app that classifies and retrains Keras models.
Created Jul 2026 · Updated Jul 2026
View Repository →Machine-Learning-with-Python
Labs of the IBM Coursera class: Machine Learning with Python
Created May 2026 · Updated Jul 2026
View Repository →Skills & Competencies
A deep skill set spanning the full AI/ML lifecycle — from raw data to deployed, production-grade models.
Tools & Frameworks
Let's Connect
Got an AI project, collaboration idea, or just want to talk ML? I'm always open to new opportunities.