Composer, poet, and computer scientist — using AI as an instrument for understanding creativity, never as a substitute for it.
I'm an artist who happens to also be a computer scientist — and the order matters. My path has been woven from music, poetry, and algorithmic thought, but it begins with making things: writing songs, setting poems, chasing the feeling of a piece clicking into place. If I had to name my real medium, it isn't music or code — it's the feedback loop between human taste and machine possibility. I make something, build a strange computational instrument to ask a better question about it, and listen to what the instrument tells me. AI is one of those instruments. It is never the author.
AI should give you more selves, not fewer. I have never wanted to automate the artist away — I want to hand people new instruments, strange mirrors, and counter-egos that send them back to their own work more alive than they left it.
My research is not about generating art more cheaply. It's a human-experience-centered inquiry: it starts from what humans actually feel and enjoy, and uses AI as a telescope to study that. Two questions drive it — what makes a song work? and how do we keep machines from collapsing into a single, repetitive voice? The first is a search for the felt logic of why art moves us (and for the places where that logic provably resists formalization). The second began when 249 of 250 poems I asked ChatGPT to write came back echoing John Keats, unprompted — a quiet loss of diversity I've been working to measure and reverse ever since, including a method (Chain-of-Specification prompting) that lets people define the kinds of difference they actually care about rather than asking vaguely for "more variety." Sameness, to me, is an aesthetic failure before it is a technical one.
As I deepen my catalog — and increasingly organize it into distinct labels, each its own aesthetic tuned for both human enjoyment and diversity — I'm building an empirical map of the space of what people find creative and moving. Understanding creativity, I've come to believe, may require producing it at a scale and variety only a human–AI collaboration can reach.
I earned my PhD in Computer Science from the University of Pennsylvania in 2024, where my work centered on making AI a personal instrument for expression rather than a homogenizing force. Today, as a researcher at Microsoft Research, I keep pressing on the same question: what genuinely enables creativity, and what shuts it down. Whether I'm demonstrating that Ableton is Turing-complete or arguing for the aesthetic seriousness of sonification, the aim is constant — empower creatives, never confine them to the limits of automation.
You can explore the research on Google Scholar, or hear the artistic side on SoundCloud.
In the age of AI, creativity is not diminished — it is reborn. Used as a collaborator rather than a replacement, technology can expand artistic expression into realms we haven't imagined, and at the same time teach us something true about creativity itself. That double promise — more art, and more understanding of what art is — is what I'm building toward.