Profile banner
Pedro Sarmento

Pedro Sarmento

AI & Music Data Scientist, Researcher, and Guitar Tablature Expert

umpedronosapato
London, United Kingdom
Joined November 2024
Favicon imageuk.linkedin.com
Favicon imagex.com
Favicon imageotnemrasordep.github.io
Favicon imagescholar.google.com
Favicon imageresearchgate.net
Favicon imageorcid.org

Network

182 connections

Summary

Pedro Sarmento is a pioneering AI & Music Data Scientist with a deep expertise in Music Information Retrieval, Deep Learning, and Music Generation. His work at companies like Moises and Music AI, combined with his academic research at Queen Mary University of London, positions him at the forefront of developing innovative AI-driven tools for music creation and analysis. nitter+1
He is a significant contributor to the academic field of AI and Music, with a PhD focused on 'Musical Smart City' concepts, exploring new ways to interpret city data through sonification. His prolific publication record includes influential papers on guitar tablature generation, symbolic music representations, and human-AI interaction in music composition, contributing to a deeper understanding of music technology. google+3
Pedro has a particular specialization in guitar-focused music information retrieval and generation, having developed significant datasets like GOAT and GigaMIDI, and models such as GTR-CTRL and ShredGP for guitar tablature and style-conditioned generation. This focus highlights his unique blend of musicality (classical guitar background) and technical AI expertise. google+2
Actively engaged in the broader AI music community, Pedro is an organizer for the AI Song Contest and frequently participates in and contributes to major conferences like ISMIR, further demonstrating his commitment to advancing the field and fostering collaboration within the domain. nitter

Work

Education

Projects

Writing

Guitar Tablature Generation with Deep Learning

January 1, 2024

Pedro Sarmento's PhD thesis, focusing on the use of deep learning for generating guitar tablatures.

Favicon imagescholar.google.com

Reflection Across AI-based Music Composition

January 1, 2024

A paper exploring reflection in AI-based music composition.

Favicon imagescholar.google.com

GTR-CTRL: Instrument and Genre Conditioning for Guitar-focused Music Generation with Transformers

January 1, 2023

Presents two conditional models for guitar-focused music generation using Transformers, with conditioning on instrument and genre.

Favicon imagescholar.google.com

LooperGP: A Loopable Sequence Model for Live Coding Performance Using GuitarPro Tablature

January 1, 2023

Introduces a method for steering a Transformer-XL model towards loopable musical structures for live coding performance using GuitarPro tablature.

Favicon imagescholar.google.com

DadaGP: A Dataset of Tokenized GuitarPro Songs for Sequence Models

January 1, 2021

Introduces a dataset of tokenized GuitarPro songs for sequence models, contributing to music information retrieval research.

Favicon imagescholar.google.com

Musical Smart City: Perspectives on Ubiquitous Sonification

January 1, 2020

Discusses perspectives on ubiquitous sonification within the context of a musical smart city.

Favicon imagescholar.google.com

Hobbies

Pedro has a strong interest in guitars and music, evident from his personal website and academic focus on guitar tablature and music generation. github

Notes

EDEmir Demirelsaid they've seen
PSPedro Sarmento
proactively help connect people who might benefit from knowing each other.

Mar 20, 2026

EDEmir Demirelsaid
PSPedro Sarmento
is one of the most impressive data scientists they know working in the music industry.

Mar 20, 2026

AEAndrew Edwardssaid they would consider
PSPedro Sarmento
a go-to expert for data science in the music industry.

Sep 30, 2025