joyce ghosn

I’m a Data Scientist specializing in machine learning, causal inference, and LLM-based agentic systems, grounded in an economics foundation. I’m currently pursuing a Master’s in Data Science at Saint Joseph University (Beirut) with study abroad experience at Université catholique de Louvain.

In my Data Science internship at NUMU Technologies, I’m building an autonomous AI architecture for quick-commerce marketing optimization that combines causal inference, knowledge engineering, MCP-based tool use, and agent evaluation. I analyze operational campaign data using panel-data methods like Difference-in-Differences and fixed effects, and I formalize domain expertise into machine-readable “skill packs” so Claude-based agents apply reasoning consistently with reduced token overhead.

I also bring applied research and business clarity from earlier work as a Business Research & Credit Analyst, compiling credit reports across the MENA region and performing ratio analysis to assess creditworthiness. My projects reflect the same end-to-end mindset—like a multimodal heart failure risk pipeline with improved test RMSE through fusion, and an LLM-driven wellness/retention agent deployed as an automated pipeline.