Matt Caren

mcaren [at] mit.edu

I'm interested in human expression—and how technology can understand, enable, and augment it.

My interests include AI for expressive music and natural language, computational cognitive science, signal processing, jazz, poetry, and musical instrument design.

I'm currently an undergrad at MIT, where I study computer science, mathematics, and music.


Research

At MIT, I develop computational models of how humans use their voices to communicate about sound. I'm advised by Josh Tenenbaum and Jonathan Ragan-Kelley, and mentored by Kartik Chandra and Karima Ma.

I also research real-time machine listening and MIR systems under Eran Egozy in the MIT Music Technology lab.

At Apple, I spent several summers developing multimodal LLM systems and bioinformatics algorithms for Apple Intelligence and the Health app.

I was also a founding member of the MIT Voxel Lab, and lead undergraduate committees on AI and interdisciplinary computing at the MIT Schwarzman College of Computing.

Publications

Sketching With Your Voice: "Non-Phonorealistic" Rendering of Sounds via Vocal Imitation  (SIGGRAPH Asia 2024)
Matthew Caren, Kartik Chandra, Karima Ma, Joshua Tenenbaum, Jonathan Ragan-Kelley

Real-time In-browser Time Warping for Live Score Following  (WAC 2024)
Matthew Caren, Eran Egozy

The KeyWI: An Expressive and Accessible Electronic Wind Instrument  (NIME 2020)
Matthew Caren, Romain Michon, Matthew Wright

TRoco: A generative algorithm using jazz music theory  (AIMC 2020)
Matthew Caren


Sound & Color

Melia

A bespoke digital instrument searching for expressiveness in the failures of AI audio-to-audio models.

> Performance

KeyWI

A next-generation electronic wind instrument.

Developed at Stanford's CCRMA, played by Grammy award-winning artists on international stages.

Etude

Original music composed for short film. Winner, Film/Media Scoring at the 2021 Marvin Hamlisch International Music Awards.

> Watch


Last updated March 2025