As complex software projects evolve, developers make a lot of changes on the source code of that project. Unfortunately, some of those changes may introduce bugs to the source code. Although, the engineers test every change before releasing it, some bugs only detected by the users. Fixing those bugs can be more expensive than fixing them before the release. Therefore, it the intention of the developers to find bugs more efficiently, and release bug-free version of their product. If they had information which changes may contain bugs during development then they could test those changes with more care. In my thesis work I investigated if it is possible to predict which changes might contain bugs, using Machine Learning algorithms. During my research as a source of data, I used the historical development data of a big scale software project called Hive. I created a system that can process the raw information stored in different sources, and I tested multiple Machine Learning algorithm on the processed data.