In this paper, we describe our ensemble approach for sentimentand aspect predictions in the financial domain for a given text. Thisensemble approach uses Convolutional Neural Networks (CNNs)and Recurrent Neural Networks (RNNs) with a ridge regressionand a voting strategy for sentiment and aspect predictions, andtherefore, does not rely on any handcrafted feature. Based on 5-crossvalidation on the released training set, the results show that CNNsoverall perform better than RNNs on both tasks, and the ensembleapproach can boost the performance further by leveraging differenttypes of deep learning approaches.