Neural networks (NN) are a subset of machine learning that has attracted great attention. However, the prerequisites of computer coding skills cause the resources for learning NN to be unequally distributed, with interested learners who do not have the background underrepresented in the subject. We propose AI4EVER to fill the gap using a hybrid of an interactive online learning environment and offline software that teaches and assists the underrepresented individuals to construct machine learning programs from scratch with no computer programming required. The learning environment has ten projects from a range of different topics that users can choose to learn. Each project has an interactive teaching module that determines if learners perform all the necessary steps and gives hints on how to proceed. Existing online resources supporting neural network customization usually have video illustrations but lack interactive learning processes. Users need to upload their data to the website, which causes data security concerns. Offline resources support users using existing models and have little support for users to construct a machine learning program from scratch. Data labeling is essential for the success of machine learning programs, but existing resources do not facilitate the process. If this project is successful, it will assist learners in understanding machine learning workflow and allow them to construct machine learning programs by drag-and-drop without prerequisite coding skills.