About Me
I’m a machine learning engineer and researcher with experience across computer vision, NLP, explainable AI, and biomedical signal processing. My work focuses on the research, design, and development of advanced machine learning systems, with an emphasis on robustness, interpretability, and real-world applicability across domains.
Currently, I work as an ML Engineer at Contrast AI, where I contribute to the research and design of generative computer-vision models for medical imaging. Previously, I was a research assistant at SRTA-City under Dr. Shaimaa Lazem, where I worked on explainable AI for NLP and evaluation frameworks for large language models. I’ve led research projects including graph-based transformer models for neurodegenerative disease classification and data-centric approaches for analyzing bias and errors in NLP systems.
more about my work can be found in my portfolio.
Interests
My interests lie at the intersection of foundational machine learning and real-world applications:
- Deep Learning & Architecture Design: Novel architectures, generative models, representation learning
- Computer Vision: Medical imaging, generative models, image-to-image translation
- Trustworthy AI: Explainability, fairness, robustness, and human-AI alignment
- Applied ML: End-to-end systems for healthcare, NLP, and signal processing
- Data-Centric AI: Systematic approaches to improve data quality and model reliability
I am driven to advance socially responsible AI — building systems that have meaningful, positive impact on society.
Featured Publications
Published
“A Data Centric HitL Framework for Conducting a Systematic Error Analysis of NLP Datasets using Explainable AI”
Ahmed El-Sayed, Aly Nasr, Youssef Mohamed, et al.
Nature Scientific Reports (2025) | DOI | Portfolio
Under Review
“A New Graph-Transformer Framework for EEG-Based Differentiation of Alzheimer’s Disease”
Youssef Tageldin, A. Eltrass, H. Farag
Journal of Neural Engineering | GitHub | Portfolio
(Preprint available upon request)
“A Comparative Framework for Evaluating Arabic LLMs in Educational Applications”
Shaimaa Lazem, Youssef Tageldin, et al.
Language Resources and Evaluation (Q1) | GitHub | Portfolio
(Preprint available upon request)
Current Work
At Contrast AI, I develop generative computer vision models for medical imaging that achieve state-of-the-art performance in clinical review and quantitative metrics. Read more: Medical Imaging AI Portfolio
Selected Projects
EEG-Based Alzheimer’s Disease Classification
A novel graph-based transformer framework achieving 99.1% accuracy for Alzheimer’s Disease classification — 7-18 percentage points above state-of-the-art
Arabic LLM Evaluation Framework
A comprehensive rubric-based evaluation framework for assessing Arabic Large Language Models in educational conversational applications
Medical Imaging AI at Contrast AI
State-of-the-art computer vision models for medical imaging applications with focus on generative models
Explainable AI for NLP Dataset Debugging
X-Deep: A Human-in-the-Loop framework using LIME and SHAP for systematic error analysis in NLP datasets - Published in Nature Scientific Reports