Tinder’s VP of Engineering has given an explainer on how the app makes use of Amazon Web Services’ (AWS) image recognition technology.
Tom Jacques, speaking at AWS re:Invent Las Vegas, says the platform’s matching algorithm uses deep-learning “AWS Rekognition” to identify key user traits from pictures. These can then be used to recommend users to one another.
AWS allows the app to mine a basic profile for personality indicators, like a guitar or travel photos.
Part of Match Group’s justification for running deep-learning on pictures in this way is that many users are abandoning the ‘bio’ section and only posting photos to their profile.
“The challenges we face are in understanding who members want to see, who they match with, who will chat, what content can we show you and how do we best present it to you,” Jacques said.
“(…) [Tinder] extracts all of this information and feed[s] it into our features store, which is a unified service that allows us to manage online, streaming and batch processing.
“We take this information and feed into our tagging system to work out what we highlight for each profile.”
The information is most clearly implemented on Tinder Picks, wherein certain profiles will be tagged “Scholar”, “Adventurer” and so on.
AWS is adding new features to its Rekognition toolkit, many of which Jacques thinks can improve Tinder’s product.
He says: “We can use this to get really deep into what is going on in our members lives and provide better services to them.”
Read more here.