I work in molecular genetics research, and I've noticed how disconnected subfields within the life sciences can be. For example, a classical geneticist working with yeast can spend their entire career oblivous to the fact that the problem they've been working on has been indirectly solved by a biochemical engineer working in the same building. This happens for a number of reasons, they publish in different journals, use different terminology, the relationship may not be obvious, etc. Originally, I only had metadata extractors and various NLP parsers specific to the bio/life sciences, but I felt that was too limiting and began to expand it. The backend which ties all the services is almost entirely written in Lua/Torch, and Redis. And everything is built around the Alfresco CMS which comes with Solr, and Mattermost as a locally hosted slack alternative. Mattermost bots report on new content (http://imgur.com/a/P3YK1, http://imgur.com/a/GTVEX) wherever it comes from.
There is too much information to stay on the bleeding-edge of things without serious commitment of resources, which start-ups don't have. My intention was to track the content a group of people go through in a day and visualize connections in the data that may not have been obvious before.
Essentially, it's meant for harvesting IP in biotech field.
