I’m Rafael Pavão, a musician, producer, and data scientist based in Portugal. I hold a PhD in Physics and currently work as a Data Scientist, applying machine learning, computer vision, predictive modeling, and statistical analysis to production systems. Alongside that, I am the co-founder of the band Devil, the Architect, where I work on songwriting, arrangement, recording, and album construction. My production approach combines musical instinct with analytical structure. I’m especially interested in shaping songs and albums into coherent sonic worlds through arrangement, recurring motifs, transitions, sound design, harmonic identity, and careful editing. My current album project, Zitterbewegung, reflects that approach: a polished art-rock record built from songs, interludes, reprises, instrumental sections, and recurring musical symbols. As a producer, I can help artists develop rough ideas into finished recordings while preserving what makes them distinctive. I’m particularly suited to alternative, art rock, experimental pop, singer-songwriter, progressive, and concept-driven projects where arrangement, atmosphere, and identity matter as much as conventional polish.…

Rafael Pavão

I’m Rafael Pavão, a musician, producer, and data scientist based in Portugal. I hold a PhD in Physics and currently work as a Data Scientist, applying machine learning, computer vision, predictive modeling, and statistical analysis to production systems. Alongside that, I am the co-founder of the band Devil, the Architect, where I work on songwriting, arrangement, recording, and album construction. My production approach combines musical instinct with analytical structure. I’m especially interested in shaping songs and albums into coherent sonic worlds through arrangement, recurring motifs, transitions, sound design, harmonic identity, and careful editing. My current album project, Zitterbewegung, reflects that approach: a polished art-rock record built from songs, interludes, reprises, instrumental sections, and recurring musical symbols. As a producer, I can help artists develop rough ideas into finished recordings while preserving what makes them distinctive. I’m particularly suited to alternative, art rock, experimental pop, singer-songwriter, progressive, and concept-driven projects where arrangement, atmosphere, and identity matter as much as conventional polish.…

Available to hire

I’m Rafael Pavão, a musician, producer, and data scientist based in Portugal. I hold a PhD in Physics and currently work as a Data Scientist, applying machine learning, computer vision, predictive modeling, and statistical analysis to production systems. Alongside that, I am the co-founder of the band Devil, the Architect, where I work on songwriting, arrangement, recording, and album construction.

My production approach combines musical instinct with analytical structure. I’m especially interested in shaping songs and albums into coherent sonic worlds through arrangement, recurring motifs, transitions, sound design, harmonic identity, and careful editing. My current album project, Zitterbewegung, reflects that approach: a polished art-rock record built from songs, interludes, reprises, instrumental sections, and recurring musical symbols.

As a producer, I can help artists develop rough ideas into finished recordings while preserving what makes them distinctive. I’m particularly suited to alternative, art rock, experimental pop, singer-songwriter, progressive, and concept-driven projects where arrangement, atmosphere, and identity matter as much as conventional polish.

See more

Language

Portuguese
Fluent
English
Advanced
Spanish; Castilian
Intermediate

Work Experience

Data Scientist at Focal Sytems
April 1, 2025 - Present
Tested and deployed an OCR pipeline for automated reading of retail product titles and price tags on shelves. Developed and productionized a full-stack planogram compliance actions tool, including backend architecture and a frontend for real-time compliance management. Analyzed visual similarity outputs to detect product mismatches and guide precision improvements. Designed a planogram-to-realogram mapping algorithm using visual similarity, vision-language models (VLMs), and an acyclical expected-order model, reaching up to 98% accuracy.
Data Scientist at SymphonyAI
November 1, 2022 - April 1, 2025
Built machine learning models and algorithms that improved neural network detection accuracy by 7%, surpassing required thresholds for a multi-million dollar deal and enabling additional agreements. Designed an algorithm to estimate shelf dimensions from mobile images to enhance inventory management. Trained an out-of-stock classification model improving F1-score by 10%, a key metric for purchase decisions. Developed reporting to track coverage of new item placement and detection, supporting optimization for a multi-million dollar customer pilot. Performed hypothesis testing on promotional displays and applied prompt engineering in language-vision models for improved object detection and contextual text recognition. Managed two junior team members, overseeing delivery of multiple ML algorithms.
App Developer & Integrator at Sonae Arauco
September 1, 2021 - October 1, 2022
Developed and optimized logistics processes using SAP ABAP and Oracle Transport Management (OTM). Conducted R&D for IT architecture design and established governance models for Azure WebApps and Power-Apps.
SAP/ABAP Programmer at Oneasit Solutions
October 1, 2019 - August 1, 2021
Designed and implemented SAP utility applications for energy consumption and customer relations, integrating REST services and handling both technical and functional components. Implemented an XML pipeline between the SAP system and an external tool for automated generation of utility letters.
Research Scientist at IFIC
October 1, 2016 - September 1, 2019
Published 14 scientific papers and presented 5 conference papers based on hadron physics research. Analyzed charmed and bottom hadronic resonances using simulations for coupling and decay-time analysis to describe previously unexplained particles and predict new ones. Developed a novel entanglement calculation method for scalar fields in a Friedmann metric background to test a cosmological constant calculation consistent with the Holographic Principle.

Education

Ph.D. in Physics at University of Valencia
October 1, 2016 - October 1, 2020
Master in Physics at University of Coimbra
September 1, 2014 - April 1, 2016
Bachelor in Physics at University of Coimbra
September 1, 2010 - June 1, 2013

Qualifications

Extraordinary Doctorate Award
October 1, 2016 - September 2, 2026

Industry Experience

Retail, Computers & Electronics, Professional Services, Energy & Utilities, Education, Transportation & Logistics, Software & Internet
    Departure

    The precursor to Sol Invictus, this album is made by a large array of musical styles. It combines inspirations from multiple musical sources, from prog rock to folk, drake ambient music, 60s love songs and more.

    It follows the story of a sailor of leaves his family in search of fortune, only to be betrayed by his crew.

    Sol Invictus

    Sol Invictus is a story-driven progressive/art-rock album built around a small number of long pieces, including the And They Became Memories / Going to Seattle / I Can Feel My Heart medley.

    It follows the story of a sailor who travels back in time to save his deceased wife, but at a great cost.

Hire a Music Producer

We have the best music producer experts on Twine. Hire a music producer in Braga today.