Participants 2025
About the TEAM
At the age of 15, I began studying guitar and took a course in computer assembly, repair, and configuration—two passions that have stayed with me ever since. I am an IBM-certified AI Engineer and also a musical data scientist, a path I’ve been developing with the help of ChatGPT. My work explores the intersection of music, programming, and data science through the creation of my own symbolic composition algorithm. Unlike typical AI-generated music, my system does not rely on machine learning or generative AI models, but rather on symbolic logic and data-driven approaches. This hands-on process has allowed me to blend my technical and musical skills in a deeply personal and creative way.
About the SONG
The song was created in the following way: segments from four different pieces generated by my composition algorithm were combined. The only alterations were removing a note from the first or second chord and trimming some notes from the final chord, leaving only the tonic (which was already present) and assigning it the full chord's duration. In other words, my role was curatorial—deciding where to place the different segments. Initially, I considered using a full song generated by the algorithm, but I couldn’t choose between two options. Eventually, I realized they were harmonically compatible and decided to create this "collage."
About the HUMAN-AI PROCESS
For the creation process, I designed an algorithm that composes music with a high level of randomness, which I consider enough to call it a “generator.” My personal touch is embedded in the design of the system, not in direct intervention with the samples or songs, but rather in knowing how to merge segments and choruses from four pieces into a single composition. The song consists of four pieces: two main ones, along with one providing the introduction and another featuring a virtuosic harp solo. I’ve been working on this system for about one or two years now.
Unlike traditional machine learning or generative AI models, my algorithm works directly with musical symbols (notes, scales, etc.). It semi-randomly triggers these musical “valves,” producing a unique composition each time. At the same time, the system preserves the internal coherence of musical harmony, ensuring that the generated sections remain connected and consonant. The algorithm even creates its own chorus during execution.
I don’t rely on generative AI models; instead, my algorithm uses symbolic programming combined with a data-driven approach. By doing this, I maintain the ability to shape the final song through my curatorial decisions, while the algorithm composes autonomously. This hybrid approach allows me to combine the system’s randomness with my artistic vision in creating the final arrangement.