At OpenTable, we help diners find the best dining experiences, wherever they travel. One of the key problems for accomplishing this is providing personalized recommendations. We have been leveraging our large corpus of unstructured reviews to build models to improve the accuracy of these recommendations. We will discuss how we use Spark both for the training of our recommenders, and for the natural language processing of the reviews to generate topic models.
Pablo Delgado is a Senior Data Engineer at OpenTable, he works on building infrastructure to collect data and power machine learning applications. Currently he is working on the recommendation systems stack using the implicit interactions of our users with restaurants. He previously worked as a Software Engineer at Google Zürich building pipelines for the Geo division. Pablo obtained a degree in Mathematics and Computer Science in University College London, London United Kingdom, where he worked on Graph based Methods for Collaborative Filtering.
Sudeep Das is a Data Scientist at OpenTable, where his main focus is on using data science techniques to enable a personalized dining experience. His current projects include applications of natural language processing and topic modeling on an extensive review corpus, to reveal salient features of restaurants and reviewers. He is also helping develop an extensive recommendation systems stack using user interactions and restaurant metadata. Sudeep used to be an Astrophysicist (Princeton PhD, UC Berkeley postdoc) researching the properties of the early universe, and co-authored about 60 papers with 2500 citations.