Seismic events such as Earthquakes are very under-researched in global scientific domains but, with advancements in Machine Learning, classification and analysis of seismic data is becoming more feasible. This project focuses on developing techniques to effectively classify Seismic data into earthquake events and seismic noise (mining blasts, man-made noise etc.). The project not only focuses on developing deep learning models but, is also focused on collecting seismic data from credible repositories with proper labelling. An extensive research in seismology ensued proper knowledge of Earthquake waves, their propagation and behavior. Main motivation behind this project is to defy already accepted triggering (STA/LTA, Z Transform etc.) as well as Picking (AR picker, baer picker etc.) algorithms and formulate new techniques to extract P and S wave features. Problem with these algorithms is that different threshold values can produce entirely different triggers and accurate thresholds are hard to generalize for large number of events.