The search team is dedicated to developing the next generation intelligent e-commerce search platform for Suning worldwide. Our mission is to deliver a pleasant online shopping experience. Whether you are shopping for a nutritious drink, a fashionable beach dress, a popular digital camera, or a getaway travel plan; in-store, on the web or through mobile app, our search platform provides an integrated, seamless and convenient user experience.
To fulfill our mission we employ state of the art machine learning techniques, natural language processing, and big data analytics to create an array of products for the eCommerce industry. Below are some of the cutting-edge projects we’re working on that will allow us to better understand and even anticipate our customers’ needs, and provide ameliorated search results to satisfy users’ query requests.
Product Word Sense:By employing sematic analysis over large amount of search, catalog and various data sources, we have developed a comprehensive lexical database of important retail product words. The database enables us to assign semantics (synonyms, hyponyms, homonyms and product category) to ‘flat’ product index and transform it into a sophisticated multi-faceted index.Word sense also enables us to intelligently analyze customer query and session intent. For example, when a user searches for a “ceiling fan,” the search engine classifies the user inputs into the “home appliance” and “air cooling functionality” category, resulting in a wider, yet specifically related scope of search results.
Query Understanding:While product (word) tagging assigns semantic meaning to individual product term, query understanding extends it further by assigning word sense to a broader collection of retail words. Functions being developed in this product include Chinese-English translation, translation from Romanized Chinese words (pinyin), identification of model number/name, as well as English and Chinese spelling corrections.. Moreover, the sematic analysis is not performed on individual search terms, but on the entire query so that we can better interpret the users’ intent.
Customer and Product Tagging:By analyzing users’ search, purchasing and consumption behaviors, we have constructed a taxonomy of online shopper persona, like “soccer mom”, “gadget geek”, “high fashioner” etc. Using these persona classifications, we are able to capture customer likes or dislikes, fashion sense, spending habits and other useful consumer information. By performing semantic and sentiment analysis on user reviews and catalog data, we are able to assign tags that characterize the products. These tags include “comfortable”, “classic Greek (home furnishing style)” and more. This allows us to personalize our search results for each individual user and, thereby, provide better user recommendations and even anticipate users’ needs.
Our team is constantly seeking collaboration opportunities with top companies and research organizations worldwide to develop a robust, elastic and high-performance search platform to websites and mobile apps for the purpose of creating a satisfying browsing experience. Our work is closely tied to our corporate ideology of making people’s search abd buying experience simple and enjoyable. Building on the open-source technologies and cloud-based infrastructure shortens our development cycles while allowing us the flexibility for customization while remaining competitive in the market.
In order to serve our ever-growing community of users, we are expanding and welcome the best and the brightest to grow with us. Here, your idea counts.