Julia has preliminary bindings to HDFS. Choose a dataset of interest to you, or one of the public datasets, and build an end-to-end analytical workflow by loading data from HDFS, analysing it in parallel, and producing some nice visualizations or plots.