News for hackers. Stuff that matters.
Re:Ask HN: Who is Hiring? (September 2011) by yummyfajitas on Thursday September 01, 2011 @10:25AM (#2950119)
Pune, India - we are looking for a general purpose developer who is smart and gets things done.

We are building a search engine/knowledge representation system for fashion.

http://styloot.com/parametric_search/search/dress-search

The best way to describe working here is to describe my last month: I built a document classifier (based on max entropy), a backbone.js app, fixed bugs in django-tyrant-cache and started building a NoSQL database which supports color-based queries ("find all the items with color similar to #df3426" [1]). Your job will be to do similar things.

(If all that stuff sounds out of your league, it sounded out of mine too, until I got started.) We use javascript, python/django, hadoop, and haskell, though we really don't care if you've never touched any of those technologies.

Contact info is in my profile. Send a link to your github or other work you've done if you are interested. Here is our github: https://github.com/styloot

[1] We aren't jumping on a silly NoSQL bandwagon. It's very easy to compute l^p distances on RGB records in your favorite SQL database, but RGB doesn't map very well to human perception of color.

Re:Ask HN: Who is Hiring? (September 2011) by nnythm on Thursday September 01, 2011 @12:35PM (#2950679)
look into mongoDB's geospatial indexing, don't reinvent the wheel.
Re:Ask HN: Who is Hiring? (September 2011) by yummyfajitas on Thursday September 01, 2011 @01:07PM (#2950801)
Mongodb spatial indexing is 2d ([x,y]), as far as I know, so that would only work if we didn't have any items with a blue channel (rgb(x,y,0)).

Regardless, that's all Euclidean. Human perception of color follows the CieLAB non-metric. (It's basically just a function of two variables, since it satisfies neither the triangle inequality nor symmetry.)

Re:Ask HN: Who is Hiring? (September 2011) by aangjie on Friday September 02, 2011 @12:00PM (#2954535)
Interesting CieLAB (Never heard of it before,but just looking through it). Function of two variables, yeah that makes sense as neurologically we have found only two types of neural cells in the retina (rods and cones).http://en.wikipedia.org/wiki/Photoreceptor_cell