The scheduling process on every production line is of principal importance as it directly affects the productivity of a company. For decades now, operational research has offered managers tools to automate their scheduling process and cope with the increase on product demand and variability. However, generally small and medium size companies do not have access to such solutions, widely incorporated on the workflow of more industrialized operations and other large companies. On this thesis, an automatic task scheduling service is developed, by implementing a GRASP algorithm with first descent local search to solve a combinatorial optimization problem known as the Job Shop Problem (JSP), along with one of its variants for considering job priorities and due dates based on total weighted tardiness (JSPTWT). This work reviews their mathematical formulation and several solution approaches and also performs a series of computational experiments to determined best set of parameters for the solver. Finally, a user interface powered by Shiny package in R is presented to offer the end users a friendly platform to support their scheduling processes, and a preliminary economic model for the first year of operation for this service is introduced.