CV

Education

• MSc in Computer Science, American University of Beirut, 2026 (expected)
• BSc in Computer Science, Lebanese University, 2024

Research experience

• Nov 2024: Graduate Research Assistant
  • Developed deep learning models for corneal tomography analysis, including U-Net and TabNet architectures for corneal disease detection.
  • Designed machine learning pipelines achieving high diagnostic performance.
  • Built 3D-CNN models for MRI analysis to study differential privacy tradeoffs and representation learning.
  • Engineered novel curvature-based indices now used clinically at AUB Medical Center.
  • Built a lightweight 1D-CNN variant caller with significantly reduced parameter count.
  • Developed a large-scale Transformer model with spatiotemporal attention, achieving state-of-the-art internal results.
  • Designed reinforcement- and imitation-learning methods for approximating convex flip distance in graph theory.
  • Supervisor: Dr. Rida Assaf, Dr. Amer Abdo Mouawad

Work experience

• Sep 2024: Graduate Teaching Assistant
  • Assisted in Operating Systems, Computer Vision, Machine Learning, and Reinforcement Learning courses.
  • Led laboratory sessions and contributed to course material development.
  • Co-developed instructional content for the AI and Data Science diploma program.
• Sep 2023: Computer Vision and Deep Learning Intern
  • Maintained and deployed multiple PyTorch-based deep learning models for urban mapping.
  • Built interactive model demos using Gradio.

Skills

• Programming Languages: Python, C/C++, Java, Bash
• Machine Learning & CV: PyTorch, TensorFlow, CNNs, Transformers
• Tools & Systems: Docker, Git, Unix, High-Performance Computing (HPC), AWS
• Other: LaTeX, IELTS (8.5)

Publications

Computational Prediction of Replication Origins: A Comparative Review of Methods, Benchmarks, and Trends from Heuristics to Deep Learning

Haidar Jomaa, Rida Assaf. (2025). "Computational Prediction of Replication Origins: A Comparative Review of Methods, Benchmarks, and Trends from Heuristics to Deep Learning." Briefings in Bioinformatics (under review).

Machine Learning-Based Detection of Keratoconus Using Zernike Polynomials Derived from Corneal Pachymetry Profiles

Haidar Jomaa, Rida Assaf, Jad Assaf, Zahi Wehbi, Shady T. Awwad. (2025). "Machine Learning-Based Detection of Keratoconus Using Zernike Polynomials Derived from Corneal Pachymetry Profiles." Poster presented at AUB Biomedical Research Days, American University of Beirut.

Raw Versus Engineered Tomographic Representations for Post–Laser Vision Correction Ectasia Diagnosis

Rida Assaf, Haidar Jomaa, Jad F. Assaf, Zahi Wehbi, Shady T. Awwad. (2025). "Raw Versus Engineered Tomographic Representations for Post–Laser Vision Correction Ectasia Diagnosis." Ophthalmology Science (submitted).

Revisiting the Inferior–Superior (I–S) Value Metric Using Machine Learning for Early Detection of Keratoconus

Shady T. Awwad, Jad F. Assaf, Zahi Wehbi, Haidar Jomaa, Rida Assaf. (2025). "Revisiting the Inferior–Superior (I–S) Value Metric Using Machine Learning for Early Detection of Keratoconus." Journal of Refractive Surgery (submitted).

Space-Time MiniLM: Spatiotemporal Attention Beyond Semantic Context

Haidar Jomaa. (2024). "Space-Time MiniLM: Spatiotemporal Attention Beyond Semantic Context." Preprint, HuggingFace.

Service and leadership

• Leadership roles in international tech and research communities (ACM, Microsoft, Huawei, NVIDIA).
• Organized and led AI-focused educational and ambassador programs, reaching 1000+ participants.
• Represented Lebanon in international academic and technical initiatives.

Honors and Awards

• MT2 Best Achievement Award, American University of Beirut
  Awarded for excellence in computer science academics and research.

• Certificate of Appreciation, Lebanese University
  Ranked first and recognized for leadership in extracurricular initiatives.