At Globant AI Studios, I built HyperShelf RetailAI, a full-stack retail intelligence platform that recovered $275.6M in stockout losses across 478 stores and generated 3.27M probabilistic demand forecasts. I also productionized a LightGBM forecasting pipeline that improved forecast accuracy by 30%.
At DAZN, I architected multi-LLM orchestration across Yellow.ai and Ada, and deployed Agentic RAG workflows that increased global ticket deflection by 35%. I developed Python middleware for Salesforce integrations and automated PII scrubbing before LLM processing.
As a Research Assistant at State University of New York, Data Science, School of Engineering & Applied Sciences, I contribute to HSAL-GNN research for sequential learning and recommendation. I also built multimodal preprocessing pipelines for cross-modal art conservation datasets, and co-authored work reported at arXiv:2506.05625.

