Radiation therapy is one of our most effective tools for combating cancers. Armed with the megavoltage x-rays generated from a medical linear accelerator, modern radiation therapy uses sophisticated optimization algorithms from a treatment planning system to calculate a therapeutic plan that delivers a lethal radiation dose to the targeted tumor while protecting the nearby normal tissues and critical structures. The implementation of modern radiation therapy starts with commissioning and quality assurance of linear accelerators and treatment planning systems. During this process, extensive machine measurements are obtained and then imported into the treatment planning system to establish an accurate model of the accelerator. After this process, the system is ready for clinical use. This is a lengthy and tedious process, where thousands of data points are collected manually by large mechanical devices. In this research, we are investigating using computer algorithmic and deep learning techniques to improve the overall process of quality assurance and commissioning of radiation therapy.