Inference in Bayesian networks is a useful tool in robotics, medicine, or other areas inwhich decision-making is prominent. It allows for informed reasoning based on what isknown about a given situation. One method to performing inference is through the use offormulas from probability theory. However, this can be inefficient and highly impractical.Automating this process using an inference algorithm allow for efficient inference that isguaranteed to terminate (assuming a reasonably sized network). This report details theimplementation of a C++ based framework for inference in Bayesian networks using thejoin-tree algorithm. A time complexity analysis is performed on the mentioned frameworkand finds that performing inference follows the expected time complexity and is in somecases about 600 times faster than a similar Matlab implementation.