If you don't know Hadoop, don't even bother applying. That's the current big-data darling, along with NoSQL DBs (esp, MongoDO and Cassandra). You should also know about distributed key/value stores (e.g., Amazon's Dynamo) and PaaS technologies like OpenStack and AWS. Be able to talk about map/reduce and know a little something about data mining or integration technologies -- companies don't usually just point a pipe at a Hadoop cluster, you'll have to write some code to comb through ERP systems or act as an endpoint on an ESB.
Fortunately, the field's young enough that there aren't too many experts yet, so if you just know the basics you should stand a good chance of getting your foot in the door. Good luck!