The High-Altitude Water Cherenkov (HAWC) Observatory comprises 300 water Cherenkov detectors, each equipped with four photomultipliers, located on the Volcán Sierra Negra in Mexico at 4,100 masl. This observatory can detect gamma rays in an energy range from 300 GeV to 100 TeV and cosmic rays from 100 GeV to 1 PeV. One of HAWC’s primary challenges is characterizing air showers and estimate their physical parameters, a highly complex task due to the nature of the data and the processes involved. Currently, HAWC employs two energy estimators for gamma rays: the ground parameter method and a neural network-based approach. However, for cosmic rays, only the likelihood-based estimator is available. In this work, we leverage machine learning techniques to achieve more accurate estimation of the physical parameters of cosmic rays. These techniques are explored as an alternative for reconstructing the physical properties of extensive air showers using simulated data aligned with the observatory’s configuration. Various models were trained and evaluated through an optimized pipeline and the most effective one was selected as the final implementation after a comprehensive comparison. This approach improves the accuracy of physical parameter estimation, contributing significantly to the detailed characterization of cosmic ray events.