Rodrigo Castellon

rjcaste [at] stanford [dot] edu

About Me

I am a sophomore at Stanford University studying computer science and mathematics. I'm currently a research assistant in the NLP Group working under Chris Donahue and Percy Liang on Music+ML, specifically building novel realistic MIDI instrument synthesizers with small data. Apart from my formal interests, in my free time I enjoy playing the piano and quickly solving twisty puzzles (Rubik's cubes), for which I had a brief competitive stint a few years ago.

At Stanford I'm also involved with the Student Space Initiative, where I serve as Guidance, Navigation, and Controls (GNC) Co-Lead on the Satellites Team. During my time working on the GNC subteam, I've helped plan and design several components that will ultimately robustly control an IBM-funded satellite in orbit launched within the next year.


Towards realistic MIDI instrument synthesizers
Rodrigo Castellon, Chris Donahue, Percy Liang
[site] [paper]

Deep Learning Control System Using an Air-Gapped Optical Input Stream

Rodrigo Castellon, William Edward Hahn
AAAS/AJAS, 2019 [code]


EduVoicer uses Deepfake/AI technology to provide personalized dictation of educational reading materials for the dyslexic/visually impaired. Top 8 Finalist, Voice Assistance Grand Prize, and Most User-Friendly Voice Assistance Hack from TreeHacks.