Adrian S. Roman

I am a Ph.D. student in Computer Science at NYU’s Music and Audio Research Lab (MARL), advised by Prof. Juan Pablo Bello, and an Audio Software Engineer at Tesla.

My research centers on developing explainable AI algorithms for multimodal machine perception, with a particular emphasis on ML/AI models for Spatial Audio Representation Learning. More broadly, I work on multimodal machine listening, including audio-visual SELD, data-driven beamforming, acoustic imaging and sound event localization and detection (SELD). I am especially interested in how auditory perception enables robust, embodied multimodal agents that can operate and reason effectively in complex real-world environments.

As a technologist, my core motivation is to translate innovative ideas into impactful products that improve people’s lives, particularly for individuals with disabilities. At Oscillo Biosciences, I developed nonlinear dynamical systems to model and simulate human synchronization, contributing to a mobile digital therapy designed to gently retrain rhythmic synchronization in patients with language disfluencies and aphasia.

As an engineer at Tesla, I develop sound-based user interfaces that help drivers intuitively navigate and interact with their vehicles. My work spans the broader software stack, including audio pipelines, UI development, and firmware, and I also build machine learning systems for speech enhancement and sound event localization and detection, bridging applied research with production-level deployment.

I hold dual Bachelor of Science degrees in Computer Science and Applied Mathematics from the University of California, Davis, and a Master of Science in Computer Science from the University of Southern California.