Adrian S. Roman

Intro

For I am involved in mankind.
Therefore, send not to know
For whom the bell tolls,
It tolls for thee.

I like to build new technologies that not only enable new paths for engineers or scientists, but also that have a postive impact on people's lives.

I am currently an Audio Software Engineer at Tesla. There, I develop audio software tools, signal processing and machine learning pipelines, and tests to ensure bit perfect audio quality in vehicles. I also develop and enhance sound user interfaces to help guide users' actions; which include, chimes and speech recognition systems.

Before working at Tesla, I interned at Oscillo Biosciences, where I had the honor of working side-by-side with Prof. Edward Large and Dr. Ji Chul Kim. There, I worked on developing dynamical systems oscillators that synchronize to music or human rhythmic interactions. I delivered an iOS app that integrates the dynamical model to re-train rhythmic synchronization on patients with aphasia.

I have a B.S in Computer Science and a B.S in Applied Mathematics from UC Davis. I am currently pursuing a M.S in Artificial Intelligence at USC. My main focus is to carry out research in machine listening: sound detection, localization and audio-visual learning. Feel free to learn more about my research.

Research

About

I was born and raised in central Mexico. My hometown, Uriangato, is a town of "doers" as they build tons of clothes used by a good portion of the nation. Despite my parents not being directly involved in clothing production, they embody the determined ethos of Uriangato, having constructed a nurturing home, built from ground up their hospital twice (the latter after their former tragically burnt), and successfully raised four children. Inspired by the prevailing "doer" mentality instilled by my hometown and family, I adopt a proactive approach to every facet of my life. Admittedly, I often find myself navigating unfamiliar territories without prior knowledge, but I embrace the challenge and learn as I go. Most importantly, immersed in the unkown of "figuring it out", I find joy seeing the world like Robert Langlands when somewhow relationships between tools and theories show up to solve a problem. Then is when I start thinking: "can I perhaps apply k-means clustering to acoustic imaging?", "can an audio-visual emotion encoder be useful to detect forgery in videos?", and the solution elegantly comes!

Contact

• Google Scholar
• Facebook
• GitHub

Elements

i = 0;

while (!deck.isInOrder()) {
    print 'Iteration ' + i;
    deck.shuffle();
    i++;
}

print 'It took ' + i + ' iterations to sort the deck.';