Data Science at Skroutz.gr

Hello folks,

It is my pleasure to announce our next event in which the team from Skroutz.gr is going to talk to us about their technology stack as well as the different challenges they have been working on. For this event we are going to visit the offices of IQbility who are angel investors and who will also give a brief introduction to their activities here in Greece. This should also be very interesting for those interested in making a startup. Our venue seats roughly 100 people but I will set the attendance limit to 150 to allow for last minute cancellations.

Besides IQbility, a massive thanks to e-food.gr who will be sponsoring food for us at this event.

Agenda

Speakers: Thalis Kapros, Nikos Fertakis

Skroutz has been the the leading price search engine in Greece for the last ten years and has steadily been expanding into new markets like Turkey (Alve.com) and United Kingdom (scrooge.co.uk). With over 4,000,000 hits per day, and over 10 million products from more than 1800 stores, there are daily management, processing and analysis needs for a high volume of data of varied nature, used by many different applications. The data management stack of Skroutz consists of both traditional and new technologies (Ruby, ElasticSearch, Mongo DB, Clojure etc.). Building recommendation systems, business intelligence tools, product classifiers and computer vision applications, the Skroutz team works daily with data trying to offer the best possible service to its users. In this presentation we will try to give an overall picture of our data management infrastructure and show how we tackle a bunch of different data-related problems in the broad area of data management (data mining, information retrieval and more).

Nikos has been a member of the Search team at Skroutz for the past 2 years. He holds a BSc in Computer Science from the University of Athens and a MSc in Machine Learning from University College London. At Skroutz he has been involved in improving the search experience for millions of users by focusing on areas such as result scoring and ranking, document indexing, auto-complete functionality and word stemming. Also, he was a member of the team that developed a data crunching pipeline for user-generated data, utilizing the ELK stack and writing a data processing layer using Clojure.

Thalis graduated from the school of Applied Mathematics and Physics of the National Technical University of Athens. He received an MSc in Computational Statistics and Machine Learning from University College London in 2011. He is a software engineer at Skroutz since September 2014, being a member of the "Earth" team, whose main scope is to handle shops' data feeds parsing and manage core operations such as products classification, analysis, and grouping. Amongst others, his interests lie in the areas of machine learning, data and text mining, recommender systems and information retrieval with a practical, engineering-based approach.