Strong experience is required in Spark.
Basic understanding of Hadoop/Big data concepts is a must.
Experience in Cloudera distribution will be an added advantage.
Should have good communication written / spoken. Expectation is member should be self-managed and able to provide effective status on individual work.
Should have at least 2 years of experience working on Hadoop eco system. Should have exposure to some of the tools available as part of Hadoop eco system. (flume / Sqoop etc). (In past we have seen that many are aware of only SQL and nothing more, hence adding this point)
It is important that the candidate is able to talk about some of the best practices or performance tips which he / she is aware of and had used in their past exp.