Introduction On the app store there are lots of mobile apps meant to help users learn about wine and teach them how to find a perfect match for the dinner. We at Onix were approached by Josselin Guibert and Julien Sahut, founders of Wine Collection Ltd., with the idea of building a unique service application, where users could make a perfect wine match while dining at a particular restaurant. Our customers came to us with the general app concept, though having most of the app’s design prepared and some supporting documentation. Josselin Guibert has become the product manager and collaborated with us on all technical aspects of development. Julien Sahut, a sommelier working with restaurants in Europe, has developed the wine <-> food matching proprietary logic. Wine Picker was created to give the answer to a simple question: How do you pick wine in a restaurant you are in and match it with your food? Approach While other mobile apps about wine offer simple wine search, the goal of our app is to help users find and make the best wine choice by using a unique-to-the-market wine selection algorithm. This algorithm performs wine selection from wine list at a specific restaurant based on wine type, rating, price, and food match, value, etc. To develop this mobile app, we have formed a solid team, where on one side we had our customers with a clear understanding of their business sphere and demand, and on the other there were the Onix team members faced with a nontrivial task of developing a complex and feature-rich mobile app that takes into account user location, wine preferences, etc. and sophisticated back-end processing of enormous amounts of data Restaurants and wine list Although the feature may seem straightforward at first glance, it implies front-end and back-end combined search, locating places, conducting analysis of place details, locating website URL, performing wine list data search on git each website and downloading wine list data to the server for further parsing. Users are also able to enrich the wine database right in the app by snapping wine lists with the phone camera as well as sending wine list files via email. Parsing the wine list Processing wine list files of various formats - .pdf, .doc, .docx, images and HTML. Image processing is two-step - first OCR engine is applied, then text is parsed for wine data. After processing the carefully screened set of wine data is further matched with wines in the DB. Pick wine The process of wine selection in a restaurant is performed fully on the front-end, where wine selection is made via a multi-step proprietary algorithm based on specific wine data and user filter options. Result We have managed to overcome the challenges, complete and release the app with much more than initially planned. As a result Wine Picker has won the acclaim at public events held. In Google search results for the UK, Wine Picker comes #1 and in search results of the app market for wine apps it is among top-10.