Yacine Ben Ameur

At Rased.ai, I lead the design and production deployment of four AI/ML products that power real-time fraud detection for financial institutions and prevent millions of USD in fraudulent transactions daily.

I build document intelligence, forgery detection, fraud-risk scoring, and natural-language-to-SQL systems, including an Arabic- and English-language document pipeline reaching 95% accuracy and an analytics agent reaching 90% single-shot text-to-SQL accuracy. I also manage four data scientists, shape model-selection and fine-tuning decisions, and contribute to product prioritization, feasibility studies, and vendor selection.

Previously, I delivered GenAI and RAG applications at Econocom, built MLOps systems and deep-learning approaches for robotic manipulation at Range Robotics, and researched reinforcement learning and uncertainty quantification for autonomous driving at INRIA.

I'm comfortable owning the full lifecycle from problem framing through deployment, using Python, LLMs, retrieval systems, computer vision, MLOps, and cloud infrastructure.