Dinh Xuan Hung

I am a researcher at the AI Security Research Center, Soongsil University. My work focuses on audio deepfake detection, spoofing-aware speaker verification, and robust speech security systems that can operate in real-world and resource-limited environments.

My recent research studies how to make audio deepfake detectors more practical: faster inference, better generalization across synthesis methods, stronger robustness to noise and domain shifts, and deployment on constrained devices. I am also interested in machine learning systems, MLOps, and tools that make research workflows reproducible.

Before joining AISRC, I worked as a software engineer at NashTech Vietnam and completed my B.Eng. at the University of Information Technology, VNUHCM. I received my M.Sc. from Soongsil University with a thesis on performance optimization for on-device audio deepfake detection.

You can find my publications on Google Scholar, my code on GitHub, and a full CV on the CV page.