Other than writing my own worker processes to pull data out of kafka (or
rather consume the messages), what frameworks/libs are there for doing this?

I know ruby has resque, and there is a java port jesque.

I don't have the server resources (or experience) setting up a hadoop hdfs
setup, so looking for a more stream lined approach.

I plan on writting my aggregate results to something like mongo/mysql.

Language of choice is java.