In this episode of the Security Repo Podcast, Maria Khodak explores how graph theory and data visualization can be used to uncover machine learning vulnerabilities like data poisoning. She explains how her work as a penetration tester intersects with research on threat detection and the importance of making abstract security concepts more human-readable. The conversation also covers her journey into cybersecurity, the value of visualization tools like Gephi, and her call to make the field more inclusive and curious-driven.Maria Khodak is a security engineer and penetration tester at a large company focusing on web application, API, and network testing. She also researches machine learning vulnerabilities, threat detection, and graph theory in her free time. https://www.linkedin.com/in/maria-khodak-10b26a192/https://gephi.org/