Recommendation engine for clothing and fashion

I want to create a recommendation engine that will be used for social shopping. The backend will need to process large amounts of data from retailer websites like Amazon, JCrew, and Gap. It will need to determine which clothing items are the most popular and then create a list of similar fashion items that can be recommended. The ultimate goal is to use machine learning algorithms to provide a tailored digital shopping experience for any user. It will be like Pandora for shopping. The user can input a link or image of a clothing item they like and the engine will recommend items to buy from other sites. This will be a long term job and our company is still in the requirements gathering phase. We are seeking excellent backend developers for this position. To be considered for this application, please attach a test program that calculates the following: (1) Create a program in php/perl/ruby that can be run on the command line. The function will calculate Nth perfect number ([login to view URL]). The input will take one argument, which is the Nth perfect number. For example, I could call your script to find the 2nd perfect number by running the following on command line: bash> php [login to view URL] 2 The result output would be: 28