PhD in Artificial Intelligence | Researcher | Educator | AI Vision & NLP Specialist.
📍 Saint Denis, France | ✉️ biljolefa@gmail.com | LinkedIn
Institution: LIPN, Université Paris 13
Co-supervisors: Prof. Hanane Azzag, Prof. Mustapha Lebbah
Topic: Digital Twin AI for Sepsis Simulation and Medical Decision Support
This PhD aims to model patient evolution via AI-driven digital twins to simulate and support clinical decision-making for sepsis. Combining multimodal learning and reinforcement strategies, it seeks to develop real-time predictive tools that enhance medical protocols through virtual testing environments.
Institution: LIPN, Université Paris 13
Co-supervisors: Prof. Hanane Azzag, Prof. Mustapha Lebbah
Topic: Multimodal and Knowledge-Driven Learning for Low-Resource Languages
This research tackles the complexity of processing languages like Kabyle. It integrates speech, text, and visual data with reinforcement learning and graph knowledge to create adaptable and robust NLP systems in under-annotated environments.
Supervisors: Prof. Mustapha Lebbah, Dr. Sylvain Gautier
Topic: Design of a Smart Communication System for Preventive Healthcare Actors
This final-year internship explores AI-based tools to enhance communication between public health professionals. The system includes intelligent search and information retrieval from research articles to support collaborative prevention strategies.
Institution: Université Paris-Saclay
Co-supervisor: Prof. Mustapha Lebbah
Topic: Molecular Property Prediction with Deep Learning
Supervised 4 M1 and 2 M2 students on tasks involving molecular solubility and toxicity using DL models. Work involved data engineering, cheminformatics, model selection and evaluation in real-world drug discovery use cases.
Institution: AIMS Senegal
Co-supervisors: Prof. Hanane Azzag, Prof. Mustapha Lebbah
Project: Extending LLMs to Vision Tasks
This project investigated cross-modal learning for tasks like object detection and phrase grounding by adapting LLMs to visual inputs, paving the way for multimodal agents.
Institution: IUT Villetaneuse
Co-supervisors: Prof. Mustapha Lebbah, HephIA
Topic: LLMs for Legal Document Processing
The project involved designing a RAG pipeline using open-source LLMs to parse legal texts. Challenges included domain adaptation and model efficiency in constrained environments.
Institution: Université Paris-Saclay
Co-supervisor: Prof. Mustapha Lebbah
Project: LLMs for Sepsis Prediction
Supervised three project groups exploring LLM-based methods to predict sepsis. Work led to the release of Sepsistral-7B, a specialized model trained on medical corpora for decision support systems.
Institution: ESP, Senegal
Co-supervisor: Dr. Khoulé
Topic: Client Profiling and Chatbot Integration
The internship focused on user segmentation for transportation services and chatbot integration for service personalization. Funded by Bakhbade Academy, it applied clustering, NLP, and deployment strategies in production environments.