Deep learning (DL) is powerful tool that has achieved significant results at challenging machine learning problems. During the last couple of years, DL has been used to address a diverse set of astronomical data analysis problems, including morphological classification of galaxies, detection of gravitational waves, transients discovery, and classification of variable sources among others. We present the basics of deep learning and show some applications in Astronomy. We particularly focus on deep learning models we are currently using within the Automatic Learning for the Rapid Classification of Events (ALeRCE) system. Our models have proven to outperform traditional methods when classifying alert streams online.