Nabil Omri

• Implementation of a multi-agent system (GenAI) for automated R to Python code migration. • Development of ML pipelines for duplicate speaker detection and merging. • Python code review, code migration, and application of best practices (Clean Code). • Preparation of technical training materials: Python, Polars, XAI.

– Designed and led the end-to-end architecture for an automated Post-Incident Report (PIR) generation system using GPT-4 and Azure. – Built multilingual transcription, speaker diarization, and key-point extraction pipelines, generating structured PIR documents aligned with AXA standards. – Delivered a review and export interface and deployed the full solution on Azure for scalable internal use.

– Designed and led an advanced RAG system for extracting clinical insights from oncology (NSCLC) research papers, including ingestion, semantic indexing, and GPT-4-based generation. – Evaluated and optimized retrieval quality using RAGAS to improve accuracy, relevance, and robustness. – Delivered a scalable Azure-based deployment leveraging Azure Search, LangChain, LangGraph, and GPT-4.

– Developed and wrote technical proposals. – Managed project engagement for a mission involving 6 developers. – Defined technical requirements for developers and facilitated communication among diverse stakeholders. – Maintained and optimized existing tools to ensure functionality and efficiency. – Developed a geographical data visualization module. – Conducted in-depth data analysis to extract actionable insights and inform decision-making.

– Developed a Deep Learning model for waste detection, designed for integration into a mobile application. – Managed and executed the raw data collection and image labelling. – Trained and optimized a custom waste detection model based on the YOLO architecture.

– Development of a method for predicting renewable energy production

– Development of a business intelligence software: web scraping & Natural language processing

– Development of an intelligent and multifunctional pupilometer: images and time series analysis

– Computer aided diagnosis of cancerous tumors – AI-based scores for metabolic syndrome screening in adolescents

– Knowledge extraction from SME data for the implementation of a prognostics and health management process. – Keywords: Prognostics and health management, Data analysis, Data quality, Statistical tools, Industry 4.0 for SMEs.

– Train operation optimization: Predictive maintenance & Operational research

– Computer aided diagnosis of cancerous tumors