Omar AbedelKader

Lille, France
Email: omar.abedelkader@inria.fr
Website: omarabedelkader.github.io
Hugging Face: omarabedelkader
GitHub: omarabedelkader
Ollama: omarabedelkader
LinkedIn: omarabedelkader

About Me

I am an AI researcher with extensive experience in artificial intelligence and software engineering.
I am currently pursuing a Ph.D. in Computer Science at the University of Lille.

Responsibilities

• Owner of the Organization Pharo-LLM
• Member of the Organizing Committee, GDR-GPL’26
• Sub reviewer at BENEVOL’25
• Project lead for ChatPharo and Pharo-Copilot
• Mentor for Google Summer of Code (GSoC)’25
• President of NeuroTech-Lille, a student association focused on AI and neuroscience
• Contributor at Ministère de la Transition Écologique (Club IA et Transition écologique)
• Contributor of the Pharo-AI Organisation
• Contributor at Conseil National du Numérique (CNNum) (Café IA)
• Supervised numerous software and AI projects at both Licence and Master’s levels, including research-oriented and applied industry projects

Publications

Experience

AI Researcher — INRIA (Villeneuve-d’Ascq, France)
Oct. 2024 – Oct. 2027
- Improving code completion and code generation using LLMs, specifically targeting the Pharo programming language, which has limited training data. - Developing techniques for code completion, type inference, and deployment in Pharo’s IDE, with a focus on runtime performance.

AI Engineer — INERIS (Verneuil-en-Halatte, France)
Sept. 2023 – Sept. 2024
- Developed “INERIS-IA,” a tool to classify textual documents according to INERIS’s strategic goals using ML and NLP techniques. - Created Boolean queries for document retrieval and improved corpus quality through document similarity and keyword extraction.

Intern — AI Researcher — LIPN (Villetaneuse, France)
June 2023 – Aug. 2023
- Compared sampling techniques for probabilistic planning, particularly for generating literary narratives. - Evaluated methods including the Score Function Estimator (SFE) and advanced techniques such as Gumbel-Softmax, assessing their effectiveness in producing coherent and creative stories.

Software Developer — BEON-IT (Beirut, Lebanon)
May 2022 – June 2022
- Java · Design Patterns · .NET Framework · Threading · Microsoft SQL Server

Full-Stack Developer — SSCC-IT (Andket, Lebanon)
Nov. 2020 – Apr. 2021
- PHP & SQL · Design Patterns · Microsoft SQL Server

Education

University of Lille — Villeneuve-d’Ascq, France
Ph.D. in Computer Science
(Oct. 2024 – Present)

University of Lorraine — Nancy, France
M.Sc. in Natural Language Processing
(Sept. 2022 – Sept. 2024)

Lebanese University — Beirut, Lebanon
B.Sc. in Data Science
(Sept. 2019 – July 2022)

Software

Pharo-Copilot

• Intelligent code completion and generation tool for Pharo, inspired by GitHub Copilot.
• Focused on repository and package -level code completion.
• Leveraged LLMs to improve developer productivity in a dynamically typed language.
• GitHub Repository: Pharo-Copilot

ChatPharo

• Live conversational interface between Pharo developers and large language models.
• Enables interactive discussions with LLMs directly from the Pharo environment.
• Designed an open and extensible architecture to experiment with multiple LLM backends.
• GitHub Repository: ChatPharo

INERIS-IA

• AI platform developed at INERIS for document classification and knowledge management.
• Built a Flask-based web interface integrating multiple ML and NLP models.
• Trained document classification models aligned with COP 2027 objectives and INERIS strategic themes.
• Implemented additional features including document similarity, keyword extraction, and corpus quality improvement.
• GitHub Repository: INERIS-IA

Family Dynamics Analysis

• This project analyzes family dynamics in France using Formal Concept Analysis (FCA).
• Using data from the French national census, the study identifies patterns in family structures, single-parent households, childbearing decisions, and naturalization through marriage across five regional zones.
• The project explores relationships between demographic variables such as marital status, household composition, and nationality, offering insights into cultural and social trends shaping French households.
• GitHub Repository: Family Dynamics Analysis
• Paper: PDF

Real–Fake Face Detection

• Demonstrates the process of building and training a neural network for classification tasks.
• Walks through data preprocessing, model architecture design, training, evaluation, and visualization of results to assess performance and prediction accuracy.
• GitHub Repository: real-fake-detection

EESMACF

• Evaluates two NLP models—SBERT (Sentence-BERT) and MiniLM—for classifying analogies using the FrameNet dataset.
• The objective is to identify valid and invalid analogies by leveraging SBERT embeddings and a fine-tuned MiniLM classifier.
• MiniLM achieved 99% accuracy in distinguishing valid from invalid analogies, outperforming SBERT (~55% accuracy).
• GitHub Repository: EESMACF
• Paper: PDF

DeGatto

• Sentiment analysis framework for e-commerce women’s apparel reviews.
• Uses a Kaggle dataset (23,000+ sentences) with aspect-level annotations for material, size, design, and comfort.
• Evaluated multiple NLP/DL/ML models (BiLSTM, SVM, Logistic Regression, Multinomial Naive Bayes). BiLSTM performed best for sentence-level analysis, and LinearSVC performed well for aspect-level analysis.
• Includes a visualization tool built with ReactJS and NodeJS to display results as bar or pie charts.
• GitHub Repository: DeGatto
• Paper: PDF

Teaching

• Teaching assistant at IUT: Introduction to Software Development 41,5 h TP

• Teaching assistant at IUT : Software Maintenance 12h TP

• Teaching assistant at Polytech: DataBase 16h TP

• Teaching assistant at Polytech: Internet 12h TP

Talks

ChatPharo: An Open Architecture for Understanding How to Talk Live to LLMs — ESUG 2025
- Slides: PDF - Video: Coming soon

Awards and Honors

Best Paper Award (3rd place) — IWST 2025, Gdańsk, Poland
“Package-Aware Approach for Repository-Level Code Completion in Pharo”

Mentorship

• Google Summer of Code (GSoC) 2025 — Mentor
• Demain (dema1n.org) — Mentor

Library

Pharo-LLM — Author and Maintainer

GitHub Repository: Pharo-LLM

• Author and maintainer of Pharo-LLM
• How to create a model for a low-resource language
• How to evaluate a model (metrics, benchmarks, validation strategies)
• Research-oriented LLM experimentation inside the Pharo ecosystem

Neural Networks Library (Pharo) — Author and Maintainer

GitHub Repository: Pharo-NN

• Native neural networks library for the Pharo programming language
• Designed and implemented core components for:
  • Network architecture definition
  • Training pipelines
  • Model evaluation
• Provides abstractions for layers, activation functions, and learning workflows
• Actively maintained and used for research and education within Smalltalk

Technologies & Skills

Programming Languages: Python, Pharo, Java, JavaScript, Machine Learning & Deep Learning: TensorFlow, PyTorch, Keras, Scikit-learn
Natural Language Processing (NLP): spaCy, NLTK
Big Data & Distributed Systems: Apache Spark, Hadoop
Data Analysis & Visualization: Pandas, Matplotlib, Seaborn
Scientific Computing: NumPy, SciPy
Web Development: Flask, Django, Laravel, ReactJS
Databases & Cloud: MongoDB, MySQL, AWS
Operating Systems: macOS, UNIX, Windows
Research Areas: Artificial Intelligence, Software Engineering

Visibility & Testimonials

Alumni Testimonial — IDMC, University of Lorraine: Featured on the official IDMC Institute website as an alumni in Natural Language Processing and Artificial Intelligence, sharing your academic and professional journey from MSc to PhD.
- Testimonial page

Volunteering

NeuroTech-Lille — President

• Led the executive board and coordinated association activities and projects.
• Managed the annual budget, including financial planning, expense tracking, and funding allocation.
• Represented the association in meetings with university administration, partners, and sponsors.
• Oversaw event planning, workshops, and conferences related to neurotechnology.
• Supervised team members and ensured effective collaboration across departments.
• Defined strategic goals and ensured alignment with the association’s mission.

Ministère de la Transition Écologique — Contributor (Club IA et Transition écologique)

• Contributed to ongoing AI-driven projects supporting ecological transition initiatives within the French government.
• Assisted in the development and evaluation of AI use cases for public policy and environmental impact.
• Collaborated with interdisciplinary teams, including policymakers, researchers, and technical experts.
• Participated in strategic discussions on responsible and sustainable AI deployment in the public sector.

Conseil National du Numérique (CNNum) — Contributor (Café IA)

• Contributed to national discussions on artificial intelligence, digital policy, and societal impact.
• Assisted in the preparation and facilitation of Café IA sessions bringing together experts, researchers, and public stakeholders.
• Participated in the analysis of AI trends, use cases, and regulatory challenges.
• Supported knowledge-sharing initiatives on responsible and ethical AI.

Croix-Rouge Française — Volunteer First Aider (2022–2023)

• Provided first aid and emergency assistance during public events and operations.
• Responded to medical emergencies following established safety and emergency protocols.
• Collaborated with multidisciplinary teams under high-pressure situations.
• Maintained readiness through regular training in first aid and emergency response.

Hobbies

• Water Activities: Swimming · Diving
• Outdoor and Adventure Sports: Hiking · Skiing · Skydiving
• Creative and Artistic Activities: Photography
• Precision Sports: Boxing · Bowling · Golf

Languages

• Arabic: Native
• French: Native
• English: B2