About me

I am a PhD candidate at the Music Technology Group of Universitat Pompeu Fabra in Barcelona under the supervision of Dr. Xavier Serra. I just finished my second research internship at Google Research Sound Understanding.

My work focuses on audio dataset creation and deep learning for sound event recognition, with a focus on learning from noisy labels and self-supervision. My thesis is partially funded by two Google Faculty Research Awards (2017 & 2018) for the creation and exploitation of the Freesound Dataset (FSD).

My most recent contributions include the release of FSD50K, which is the new open dataset of human-labeled sound events with over 100h of audio, and a paper accepted for ICASSP21 about unsupervised contrastive learning of sound event representations. Currently, I’m finalizing a paper with the results from my recent second internship at Google Research Sound Understanding.

I have been a Task organizer of the DCASE Challenge for a couple of years - the last one was DCASE2019 Task 2 Audio tagging with noisy labels and minimal supervision, run on Kaggle with 880 teams participating.

My research interests include machine learning and signal processing with applications to machine listening, as well as sound and music computing and acoustics.

I have also been involved in the teaching of one undergraduate course (Sound and Music Processing) and in Coursera’s Audio Signal Processing for Music Applications as teaching assistant. I graduated from Technical University of Madrid (Spain) in Telecommunications Engineering and received my Master’s degree in Engineering Acoustics from Aalborg University (Denmark).

In this web you will find information about me and my work, the research that I do and some other possibly interesting stuff.

You can also reach me on: