Aymen Adjlane

– Reinforcement learning (RL) with graph neural networks (GNN) for anomaly detection. – Technologies used: PyTorch, PyTorch Geometric, Numpy, Pandas.

Added bilingual support (Arabic and French) to a startup's platform, with corresponding CSS style adaptation. Technologies used: HTML/CSS, NodeJS, ExpressJS, i18next, Handlebars.

– Design and development of a chatbot based on large language models (LLM) specialized in law. – Exploration and comparison of several models: CohereForAI/aya-23-8B, AceGPT-7B (based on LLaMA2), GPT-4o-mini. – Experimentation with different RAG techniques: Query Routing, HyDE, Multi-Query, Re-Ranking. – Technologies used: HuggingFace, LangChain, OpenAI APIs, ChromaDB, TruLens.