Mohammed-Amine BAGUENNA

• Build an end-to-end architecture of an AI-driven Job Descriptions to CV semantic-matching platform using LangChain, Google Gemini API, ChromaDB,FastAPI, River online-learner fusion. • Built a real-time feedback loop allowing continuous relevance tuning and raising match precision. • Deploying the stack in Docker Compose (watcher, parser, matcher, learner, UI) with GPU-ready images. • Present best-fit CVs in seconds, saving recruiters days of manual screening

• Web Scrapping (Selenium, Web driver) and Data Preprocessing (Pandas, Numpy). • Fine-Tuning Masked Language Model (MLM) for Moroccan Darija linguistic specificities (Transformers, LoRA, QLoRA). • Generating text embeddings for downstream NLP tasks.

• Web Scrapping (Selenium, Web driver, Beautiful Soup) and Data Preprocessing (Pandas, Numpy). • Creating an OCR (Optical Character Recognition) for Arabic characters. • Improving OCR’s results used by the company. • Designing of a machine Translate model (French – Arabic).