As a part of the capstone project at Carnegie Mellon I got the opportunity to work with Verizon Business and use different machine learning techniques to calculate risk and confidence score for Indicators of Compromise (IoC) and classify it as malicious, suspicious, or benign. We also created a data pipeline consisting of a JSON Parser to bring raw data in usable format as well as a feature extractor to derive valuable features from each IoC to boost model performance. The following video gives a detailed overview of the project. You can also view the video here
Feel free to get in touch for more details on code and methodology used