In this talk, Tatiana will focus on the detection of malicious code in machine learning models. She will talk about the structure of file formats used for storing such models, the process of running them, and the existing attack vectors. The speaker will also present the results of a study on popular community repositories