Justin Salamon
Principal Scientist & Research Manager at Adobe
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolJustin Salamon is a Principal Scientist and Research Manager leading Adobe Research’s Sound Design AI Group (SODA) in San Francisco, where he builds machine learning and signal-processing systems for sound generation and audio–visual creative tools. With 12 years of experience spanning academia and industry, he applies representation learning, self-supervision, and multimodal machine listening to problems from music information retrieval to bioacoustics and audio-for-video. He has a strong research pedigree—PhD in Sound and Music Computing—and a track record of shipping open-source tools and evaluation code used by the MIR community, including contributions to the widely cited CREPE pitch estimator and mir_eval evaluation suite. Known for bridging rigorous research with practical engineering, he focuses on usability and robust APIs as much as model performance. Based in San Francisco, he combines academic publication depth with product-minded delivery at scale.
12 years of coding experience
14 years of employment as a software developer
BA Computer Science, BA Computer Science at University of CambridgePhD Sound and Music Computing, PhD Sound and Music Computing at Universitat Pompeu Fabra
English, Spanish, Hebrew, Catalan
Github Skills (19)
algorithm10
algorithms10
python10
evaluation10
command-line-interface10
implement10
numpy10
music-information-retrieval10
command-line10
scientific-computing10
metric10
cli10
signal-processing9
interpolation9
tensorflow9
Programming languages (7)
Github contributions (5)
mir-evaluation/mir_eval
Mar 2014 - May 2016
Evaluation functions for music/audio information retrieval/signal processing algorithms.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:130 commits, 4 PRs, 77 pushes in 2 years 2 months
Contributions summary:Justin contributed significantly to the implementation of a melody extraction evaluation module. They developed core functionality, including resampling melody sequences, calculating various evaluation metrics (voicing recall, false alarm rate, raw pitch, raw chroma, overall accuracy), and handling unvoiced frames. They also refactored the code into a separate evaluator and implemented functions for individual evaluation measures. The contributions involved significant changes to the melody.py file, adding new capabilities for assessing the performance of melody extraction algorithms.
signalmlsevaluationaudiosignal-processing
marl/crepe
Apr 2018 - May 2020
CREPE: A Convolutional REpresentation for Pitch Estimation -- pre-trained model (ICASSP 2018)
Role in this project:Full-stack Developer
Full-stack DeveloperContributions:3 releases, 31 commits, 10 PRs in 2 years 1 month
Contributions summary:Justin primarily worked on enhancing the CREPE pitch estimation script. They improved the command-line interface by updating arguments and descriptions, and introduced more robust API functionalities. The user made consistent code formatting improvements and added support for padding and other related functionalities. Their contributions demonstrate a focus on the project's usability and overall codebase quality.
pytorchrepresentationpre-trained-modelpre-trainedconvolutional
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.