Anneswa Ghosh is an Applied Scientist at Microsoft AI, where she works on the Shopping team building intelligent consumer experiences across Microsoft Edge, Bing Search
and Copilot. She has contributed to innovative AI-powered shopping features, including coupon savings experiences in Edge and shopping recommendations in Microsoft’s consumer Copilot. Her work combines strong applied research depth with production execution, particularly in machine evaluations, recommendation quality, ranking systems, and building 0-to-1 recommendation systems that help users make better purchasing decisions.

Anneswa’s research background spans deep learning, large-scale prediction systems, privacy-preserving AI, and responsible model evaluation. Her master’s thesis at the University of Utah, Spectrum Usage Analysis and Prediction using LSTM Networks, applied LSTM-based deep learning methods to wireless spectrum usage prediction, studying frequency utilization and identifying under-utilized bands with predictable usage patterns. More recently, she co-developed PANORAMA, a large-scale synthetic dataset of 384,789 samples from 9,674 realistic human profiles, designed to model how Personally Identifiable Information (PII) and sensitive attributes appear in online content. This work provides researchers with open-source resources for auditing privacy risks, evaluating sensitive data memorization, and developing mitigation strategies for safer LLM deployment.