Now these points of data make a beautiful line
And we're [launching] beta; we're releasing on time.
Well, it's official: I've been working the past year at Evri, and we've just launched. Basically the system leverages natural language parsing to create a rich data graph of all the concepts on the Web that we can get our hands on. I could go into more detail, but the official post linked above does a great job of that. It's pretty cool to browse around and see the article recommendations for a given entity. Unfortunately the visualization aspect requires the Flash plugin, so if you don't want to put up with the insane crashes, UI glitches, and performance issues that come along with installing that plugin then you miss out on that goodness. Hopefully Gnash will be able to run it soon.
Anyway, it's great to have something public and at the same time useful, especially after the how my last few jobs have fared. My role in the system has been in content gathering rather than the core parsing system. It's been interesting mostly simply because I haven't had to work with data sets in the bajilli-byte range before, and you run into a whole new set of challenges dealing with that kind of thing.
I've also been able to use and contribute to a number of free software packages, which has been a big plus. This is the first time a work project has really pushed development of new features in another library rather than just responding to bug fixes, so it's been great to get involved with the hacker community on work time.
Anyway, give it a look; poke around; kick the tires. I bet you'll find something interesting.