Abdelrhman
Kabbary
Data Scientist & Machine Learning Engineer
Building end-to-end ML pipelines, statistical models, and data-driven solutions across banking and telecommunications. Transforming complex datasets into actionable business insights.
Expertise
Technical Skills
Proficient across the full ML stack — from data wrangling and statistical analysis to deep learning and generative AI deployment.
Career
Work Experience
Professional journey spanning ML engineering, data analysis, and IT systems across diverse industries.
- Developed and deployed supervised ML models applying end-to-end pipelines from data ingestion and feature engineering to model evaluation and deployment.
- Conducted EDA on real-world datasets, uncovering patterns and delivering actionable insights to guide model design decisions.
- Implemented deep learning architectures (CNNs, RNNs, Transformers) using PyTorch and TensorFlow, focusing on performance optimization.
- Collaborated on generative AI projects utilizing LLMs, applying prompt engineering and fine-tuning techniques to solve domain-specific problems.
- Completed structured program covering AI in Banking, ESG frameworks, Green Finance, Cybersecurity, and Ethical AI.
- Examined real-world applications of AI/ML in banking — intelligent personalization, fraud detection pipelines, and operational automation.
- Analyzed ESG risk assessment frameworks and responsible investing principles, understanding how sustainability metrics shape financial strategy.
- Explored cybersecurity and data privacy practices in financial institutions, reinforcing model risk management awareness.
- Completed intensive training in data analysis, statistical modeling, and ML fundamentals applied to telecom-domain datasets.
- Performed data cleaning, transformation, and EDA producing statistical summaries and visualization reports.
- Implemented regression and classification models in Python (Scikit-learn) evaluating with accuracy, RMSE, precision, recall, and F1-score.
- Built SQL queries for data extraction across relational databases using joins, aggregations, and window functions.
- Administered and maintained Odoo ERP, managing user accounts, RBAC, and data governance policies across departments.
- Organized, audited, and reconciled financial records improving data accuracy for operational decision-making.
- Implemented a smart membership card and NFC sticker system enabling automated member identification at entry gates.
- Managed end-to-end data pipeline for member records ensuring accurate data entry and real-time retrieval at entry points.
Portfolio
Featured Projects
End-to-end data science and ML solutions spanning NLP, computer vision, time series, and generative AI.
End-to-end pipeline from automated web scraping of major used car marketplaces, through EDA and an interactive Power BI dashboard, plus an ML regression model for price prediction.
ML pipeline for telecom churn prediction (7,000+ records). Benchmarked Logistic Regression, Random Forest, and XGBoost — achieving 88% accuracy and 0.91 AUC-ROC with Power BI dashboard.
Time series forecasting for monthly sales trends using ARIMA and Facebook Prophet. Stationarity testing and decomposition achieved MAPE below 6% with visual confidence-interval reports.
Fine-tuned BERT transformer on 50,000+ customer reviews achieving 92% accuracy. Full NLP workflow: text cleaning, tokenization, embedding generation, training, and deployment as a lightweight inference API.
End-to-end RAG pipeline for natural language Q&A over legal contracts. SentenceTransformers + FAISS retrieval, multi-layer safety guardrails, LLM-as-a-Judge evaluation, and streaming FastAPI + Gradio UI.
↗ View on GitHubComprehensive EDA on banking transactions dataset with 15+ features — statistical distributions, outlier detection, correlation analysis, anomaly flags, and customer segmentation recommendations.
Academic
Education & Languages
🌐 Languages
🏆 Certifications
Azure AI Services, Cognitive Services, NLP
Cloud AI & Machine Learning on Azure
Stanford & OpenAI (Andrew Ng)
ICT Developing Professional – AI
DataCamp — PostgreSQL (ID: SQA0016829178916)
NVIDIA & ITI
QPoland (Diploma No. QML Workshop1-105)
DeepLearning.AI & Neo4j
Hugging Face Agents Course
DataCamp
Community
Volunteer Work
Contributing time and skills to cultural and scientific institutions in Alexandria.
- Captured and edited event and promotional videos for the International Book Fair.
- Created social media content and collaborated with the media team.
- Completed 129 volunteer hours.
- Coordinated stage activities for talks, shows, and live sessions.
- Managed schedules and assisted renowned speakers and public figures during sessions.
- Ensured smooth event execution by coordinating with the organizing team.
Get In Touch
Let's Connect
Open to data science roles, ML engineering opportunities, and collaborative projects. Let's build something great together.