Tailor made stores at myntra or how to personalize your search results
Submitted by Apoorva Gaurav (@apoorvagaurav) on Saturday, 31 May 2014
Section: Crisp talk Technical level: Intermediate
This will showcase a unique way of personalization which is a combination between search and recommendations. Here we'll not go into the details of the algorithms of how a product is deemed suitable for a user but given that the product has been shortlisted on certain criteria how to show case that.
The talk will give some idea of Cassandra and Solr.
Myntra is India's biggest online fashion store. We sell hundreds of thousands of apparels and footwears hence we continuously face the problem of how to showcase our vast catalogue and how to make sure that a user gets to the prodcut (s)he is looking for with minimum hassle. We've tried various approaches with varying degree of success. The approach we've taken is kind of non-intrusive cross between recommendations and search where on any product listing page we are first showing products based on user past behaviour and then the usual products.
Associate architect at myntra with over 10 years of experience. Dabbling in big data and search related problems for last four years.