[ganeshkrishnan]

>Search relevance is a serious high-science research problem

We are working on this scenario at http://www.shoten.xyz using document clustering and apache spark graphx+ giraph.

Ain't been an easy task so far but we have made some headway