During this talk, I will discuss Gousto's journey (both the highs and the lows!) to implementing our recipe recommendation system. Being able to recommend the right recipes is a crucial part of offering a convenient service and we have therefore implemented several approaches. I will go through the different recommendation methods and the learnings we have taken from each of them.
With Gousto offering up to 40 recipes a week, we believe that recommending the right recipes to our customers is key to being able to provide them with a convenient service and meaningful choice.
To this end, we have implemented several recommendation solutions ranging from a purely content based recommender, using Python and graph database neo4j, to a collaborative method which leverages clustering. During the talk, I will discuss the lessons we have learnt along the way, the results we have achieved and also touch on how the recommendation engine is integrated into the wider business.