We're using this at my work for a product recommendation engine. It's essentially a wrapper around Apache Mahout (which uses Hadoop). It make the whole Hadoop/M

We're using this at my work for a product recommendation engine. It's essentially a wrapper around Apache Mahout (which uses Hadoop). It make the whole Hadoop/Mahout setup much more accessible, but it still has the same drawbacks (sucks memory like anything, lots of overhead to ramping up the jobs)