At the New York Times Interactives desk, where I was an intern in the summer of 2012, I and two others built an internal web app, named Imago, for processing and curating Olympics photos.

Imago was built in Ruby on Rails and Backbone.js, and relied heavily on Resque. The app received (via FTP) up to 1000 photos per hour from the Times’s own photogs and from the wires. These photos were cropped and their captions processed by an entity-extraction engine to deduce which athletes, sports, countries and events were referenced in the caption. These photos were then surfaced in the Backbone app which allowed a photo editor to filter the collection of photos across a variety of (composable) facets, edit photos’ tags and captions, rate photos and create slideshows.

JSON representations of the various facets (i.e. soccer photos or U.S.A. photos) were baked out to S3 and served as the backend for this photo explorer (and a handful of others). Slideshows built by Imago were featured on the front page of multiple times during the course of the Olympics.