Machine Learning Engineer, Capital One
Kate Highnam was recommended for this list based on her presentation on "Deep Learning for Realtime Malware Detection" at SchmooCon 2018, where she focused on command-and-control activity from domain generation algorithms (DGAs). She and her colleague, Domenic Puzio, used deep learning to understand patterns of domains that malware will likely use for command and control, allowing them to detect malicious websites and machines. The presentation was "at the Ph.D level," according to one attendee.
At Capital One, Highnam uses machine learning for advanced malware detection. "Part of what makes Kate unique is that she has both cybersecurity knowledge and machine learning expertise, so she is able to use these advanced learning algorithms to solve really complex security challenges," says Puzio.
She also has an engineering background to take a machine learning model and productionize it at scale. As a University of Virginia student, Highnam's thesis was a published industrial research paper which included an attack scenario and repair algorithm for drones on missions with limited contact from ground control.
Her current research project investigates beaconing patterns in large-scale network traffic. She and Puzio have a publication under review for their DGA research and will present a keynote at this year's Deep Learning World called Realtime Malware Detection with CNNs and LSTMs.
Image Source: Kate Highnam