Early identification of areas that are at-risk in a design under development can help design and verification teams take early preventive and corrective measures. We propose a novel system that predicts which files are at risk of being buggy. Our simple, machine learning-based, regression system receives data from various sources, including the version control and defect tracking tools. After automatically labeling historic data and training sets, the system predicts the risk level of each file. Experimental results show that the proposed system has fairly high positive precision and recall and that its predictions are better than trivial predictions, such as files that recently changed or had bugs.