Ugur Demirkaya

Data Scientist with 5+ years building production AI systems — from XGBoost forecasting engines and BERT-based RAG chatbots to MILP-driven optimization platforms. Specialized in end-to-end pipelines: feature engineering, model training, MLOps (Docker, MLflow, FastAPI), and decision science (OR-Tools). Proven impact: 30% reduction in SLA breach risk, 50% faster information retrieval, automated lab resource scheduling via mathematical optimization. Currently pursuing an MSc in Data and Information Engineering while developing LabAI_Optima (lab optimization combining OR + ML) and a multi-layer Fraud & Anomaly Detection Platform