United States
AI expert specializing in machine learning and NLP to solve complex industry problems at scale.
Develop an intelligent revenue cycle management (RCM) model to predict the likelihood of claim rejection. Utilize natural language processing (NLP) techniques (such as Transformer models) to extract medical coding information from tens of millions of noisy patient medical records (such as scanned images). Use image processing techniques to detect dental diseases such as tooth decay and bone loss. Simultaneously, combine graph neural networks (GNNs) with knowledge graphs (KGs) to assist hospitals in disease prediction, surgical indication recommendation, and image report interpretation.
I have led several projects aimed at enhancing the Expedia travel experience, spanning multiple areas. Projects currently deployed or under development include: chatbots trained on deep neural networks, supporting multiple business units such as customer service, VRBO listing search, Expedia hotel search queries, and hotel supplier (or host) Q&A. These chatbots cover multiple countries and support multiple languages; a question-answering engine trained on deep neural networks, capable of extracting answers from unstructured customer reviews, existing answers, and other unstructured content; an accommodation knowledge graph connecting all aspects of travel, quickly generating personalized recommendations during traveler searches; a market bidding strategy based on reinforcement learning, designed to maximize the ROI of paid advertising; a hotel pricing model based on global accommodation market demand forecasts; and end-to-end itinerary recommendations based on traveler preferences to generate personalized travel suggestions.
Develop personalized hotel search ranking models for business travelers to enhance their hotel search experience. Simultaneously, build a rating engine to rate hundreds of thousands of hotels globally based on payment behavior. Establish a global partner payment analytics infrastructure and lead multiple functional teams in analytics efforts to drive the multi-billion dollar Expedia Virtual Card program. Take effective measures to attract more hotels to participate in the program. Furthermore, leverage data science for advanced analytics to predict booking growth and support strategic decision-making by senior management.
By leveraging integrated machine learning techniques and simulation methods, we can predict freight volumes within the Amazon FBA network with extremely high precision, capturing dynamic changes. Simultaneously, we process and analyze tens of thousands of highly correlated and volatile time-series data points daily, combining predictive and optimization models to forecast constrained yet relevant time-series data. As the Head of Forecasting for Amazon Europe Operations, I am committed to driving operational excellence by extracting valuable insights from massive datasets through accurate forecasting and in-depth analysis.
During my time as a research assistant in the Machine Learning and Data Mining Group at Microsoft Research Asia (2010-2011), my research focused primarily on Natural Language Processing (NLP), Computer Vision (CV), and multimodal representation learning. I participated in the prototype development of a search recommendation system based on topic modeling and deep neural networks, providing Bing Search with context-aware content recommendation capabilities; and I studied early convolutional neural network (CNN) models for image semantic understanding, laying the foundation for the application of transfer learning in subsequent visual tasks.