Rôles occupés
Back-end
Miscellaneous
ETL Datawarehouse
Front-end
Conducted a comparative study of LLMs (BERT, Phi-3, Mistral, Gliner) for Named Entity Recognition (NER) and medical Question Answering. Implemented RAG pipelines to improve information retrieval and contextual understanding. Integrated medical knowledge graphs (Neo4j) to enrich prompts and enhance analysis of complex entities. Designed a qualitative and quantitative evaluation protocol for NER and QA performance. Deployed backend (FastAPI) and frontend (React) applications with a CI/CD pipeline for automated LLMOps workflows.
Developed Building Gene, an AI-powered real estate sales platform for instant lead qualification and engagement. Designed AI agents for CRM automation (Relevance AI, OpenAI, SMS, WhatsApp). Built an intelligent voice assistant (VAPI) for call handling and customer service automation. Optimized multi-agent workflows with Pipedream and GoHighLevel.
Conducted benchmarking of LLMs for code generation and optimized fine-tuning and inference workflows. Designed an evaluation protocol to ensure code quality and updated databases according to performance outcomes. Integrated a RAG architecture to improve chatbot contextualization and relevance for QA processes.
Developed an ETL pipeline (GCP BigQuery) for ingestion and processing of multimodal economic datasets. Built real-time interactive dashboards with Power BI, Plotly, FastAPI, and MongoDB. Applied ML models (clustering, classification) to analyze Tunisian import/export flows and identify commercial opportunities.
Engineering Degree in Data Science.
Preparatory Scientific Cycle in Mathematics & Computer Science.