✧ AI Engineer with 2+ years of hands-on experience in machine learning, deep learning, and data science, coupled with a strong foundation in software development. Proven ability to design, build, and deploy scalable ML/DL models using tools such as Python, TensorFlow, Keras, PyTorch, scikit-learn, and Pandas. Experienced in developing end-to-end pipelines for data preprocessing, modeling, evaluation, and visualization.
✧ Adept at collaborating across teams to deliver impactful, AI-driven solutions in real-world applications. Demonstrated success in implementing models for classification, object detection, time-series prediction, and data imputation.
✧ Passionate about continuous learning, adopting best practices in model development and MLOps, and contributing to innovations that drive business value and user experience.
ALRIFAI CONSULTING GROUP | Istanbul, Türkiye
October 2024 - Present
Collaborated with a multidisciplinary team under experts supervision to conduct applied AI research. Played a key role in authoring a research paper titled "Strategic Measurement in the Age of AI: Improving KPIs for Construction Companies: Evidence from Saudi Arabia" published in a globally recognized Q2-ranked journal. Contributed to the design of experimental frameworks, data analysis, and AI model evaluation. Demonstrated strong academic profeciency and a commitment to real-world impact through scientific research.
Major of Artificial Intelligence and Natural Languages Processing
Damascus University | 2025 - Present
Currently pursuing a Master's degree in Artificial Intelligence and Natural Languages Processing at Damascus University.
Major of Artificial Intelligence and Natural Languages Processing
Damascus University | 2019 - 2024
Graduated with a comprehensive understanding of AI principles, machine learning algorithms, and software engineering practices.
Python
Java
C++
Shell Scripting
TensorFlow
PyTorch
Pandas
Scikit-learn
Docker
Git
Terraform
Airflow
Kestra
MySQL
PostgreSQL
Google Cloud Platform
dbt
Excel
Spark
Power BI
Matplotlib
Developed an advanced system using Autoencoders, LSTMs to accurately reconstruct incomplete motion-capture data, managed to Reconstruct cases were 30 out of 41 markers were missing with less than 5cm error, crucial for Animation and Biomechanics.
Technologies: Python, TensorFlow, Keras, Pandas, NumPy, Matplotlib, BTK (Biomechanical ToolKit), os, Gradient Clipping & Weight Decay, Noise Injection, Masking Strategy, binary data handling.
Led a comprehensive data science project to analyze and model traffic accident patterns in Chicago using open-source crash report data. The goal was to uncover key factors contributing to leading cause of accidents, Accident severity , identify spatial and temporal trends, and apply machine learning techniques to enhance predictive insights.
Technologies: Python, Pandas, NumPy, Scikit-learn, SciPy (Hamming, Jaccard, Cosine, Minkowski distances, Chi-Square Test), GeoPandas, Shapely, Matplotlib, Seaborn.
Designed chatbot application Diagnoses diseases and suggests treatments, Provides medication and dietary advice, Performs menstrual and water analysis. Provides disease information sourced from FDA and PubMed, Includes plant-based treatment options with cautionary advice, Delivers personalized health insights based on user data
Technologies: Python, NumPy, Pandas, Scikit-learn, TensorFlow Keras (Tokenizer, pad_sequences, Sequential, Embedding, LSTM), Flutter, Laravel.
Led a comprehensive data science project to analyze and model student dropout patterns using educational institution records and demographic data. The goal was to identify key factors contributing to student dropout, assess the impact of socio-economic and academic variables, uncover temporal trends, and apply machine learning techniques to predict dropout risk and support early intervention strategies.
Technologies: Python, pandas, numpx y, scikit-learn, scipy, matplotlib, seaborn.
An end-to-end automated jigsaw-puzzle solver. It starts with image preprocessing (piece extraction and contour detection), feature representation (color and shape descriptors), similarity scoring (Dynamic Time Warping), and puzzle assembly (greedy matching and high-level filtering). The project leverages OpenCV for low-level vision, NumPy for data handling, and classic algorithms (DTW, greedy search) to achieve fast, accurate piece matching and assembly
Technologies: Python, OpenCV, NumPy, DTW, greedy search.
.png)
.png)
.png)
.png)
Developed a robust machine learning model to predict real estate prices using advanced regression techniques on structured housing data. The project involved end-to-end pipeline design including data cleaning, feature engineering, and model tuning with a focus on interpretability and performance. I built a predictive system that accurately estimates property values based on key features such as location, size, and amenities. The model achieved high predictive accuracy, with a low Mean Squared Error (MSE) and a strong R² score, demonstrating its effectiveness in generating actionable insights for real-world decision-making.
Technologies: Python, pandas, scikit-learn, seaborn.
This project addresses the challenge of spelling correction in the Arabic language by leveraging advanced Natural Language Processing (NLP) techniques. Due to the morphological complexity and contextual ambiguity inherent in Arabic, traditional spelling correction approaches often fall short. My system proposes a context-sensitive solution using Bidirectional LSTM-based language models to detect and correct both real-word errors (RWEs) and non-word errors (NWEs) in Arabic text.
Technologies: Python, TensorFlow, NumPy, NLTK (for n-grams and tokenization), Other necessary libraries like re, string, and sys
This project focuses on the development of a hand gesture recognition system using state-of-the-art deep learning techniques. Hand gestures play a crucial role in human-computer interaction, offering a natural, intuitive, and touch-free control mechanism for digital systems. The system is powered by a Convolutional Neural Network (CNN) trained to classify a variety of static and dynamic hand gestures with high accuracy. The model is optimized for real-time performance and is suitable for integration into applications such as: Gesture-based control systems Augmented Reality (AR) and Virtual Reality (VR) interfaces Assistive technology for users with physical impairments With a strong emphasis on accuracy and responsiveness.
Technologies: Python, TensorFlow, Keras, OpenCV, NumPy, Matplotlib, Seaborn, scikit-learn.
The project embodies a highly intricate pursuit within the field of astrophysics - a precise physical simulation that accurately replicates the complex motions of celestial objects. This simulation provides an all-encompassing representation of gravitational forces, illustrating the mutual attraction between these celestial entities while considering fundamental aspects such as mass and diverse physical attributes. We can add bodies to the scene, aiding in predicting astronomical events and deepening our comprehension of cosmic dynamics. Moreover, The simulation offers the capability to manipulate time, allowing for accelerated progression and predictive analysis of future celestial phenomena. In celestial systems, F = G(m₁·m₂)/r² explains how a massive body like a star or planet exerts a dominant gravitational force.
Technologies: Three.js, JavaScript, Html, CSS
.png)
3D Angry Birds remake in Unity using a bottom-up architecture with minimal reliance on Unity's built-in physics. Features include advanced impact modeling, precise collision detection, structural fracture, explosions, object deformation, projectile pulling, strategic bird launching, and trajectory prediction. All classic bird types (normal, heavyweight, split, explosive, high-velocity) are included, along with material simulation for glass, wood, and iron, and force-based damage for realistic destruction and pig elimination. The project uses BVH, AABB, Swept AABB, and SAT algorithms for fast, accurate collision detection, This combination can handle up to 13 million triangles and 10 million vertices , With Unity Mesh Simplifier for performance; Mesh Destroy for realistic explosions; and custom scripts for bird abilities. These algorithms manage physics, collisions, performance, and gameplay effects.
Technologies: Unity Game Engine, C#
Contributing to the community through knowledge sharing, mentorship, and open-source collaboration.
Vienna, Austria - May 2024 - Present
I began my journey at Reparametrize as an AI Researcher, where I contributed to applied research in artificial intelligence and developed solutions for real-world challenges. Over time, I expanded my role by working on several AI-driven projects. Currently, I serve as the Co-Head of the IT & AI Department, where I lead a team of engineers and researchers. My responsibilities include overseeing AI and IT initiatives, driving innovation, managing technical strategies, and ensuring the successful delivery of projects that align with the organization's vision.
I am always excited to discuss new projects, innovative ideas, or potential collaborations. Feel free to reach out directly or use the form below.