Recommendation engine plays an important role in user experience. Most of the giant companies like Amazon, Google, Facebook try their best to personalize the recommendations for each customer. For example, YouTube gives you personalized recommendations based on your watch history. Amazon gives you a different set of recommendations like “Related to items you’ve viewed”, “Customers also bought”, “Items you may like”,.etc. Amazon sends you personalized emails based on your browsing and purchasing history. Most of the companies have their own recommendation engines to increase their customer’s conversion rates. Giving personalized recommendations in real-time is very very difficult. For big companies, the user base and the item catalogue will range from millions to billions. Handling such huge amount of data and recommending a user in real-time is a nightmare. In this article, I will summarize the paper “Pixie: A System for Recommending 3+ Billion Items to 200+ Million Users in Rea
Data Scientist | Tech Blogger