PROJECTS
Bewildering Brain
An exploration of Markov Chains and RNNs (Recurrent Neural Networks), a comparison and analysis of which one outputs better results and how they both perform when predicting the lyrics of Bob Dylan (or Luis Alberto Spinetta in the Spanish version). Pulled lyrics from Genius API and guided myself with Spotify’s Web API.
Spotify ReWrapped
Pulling data from my Spotify and Last.fm accounts using their APIs, I did some time series analysis to check which were my real stats and see if what Spotify told me in my Spotify Wrapped end-of-year summary was correct. Also did some interesting data visualization to help explain my insights.
Data, data
An immersive analysis into Jorge Drexler’s universe through statistical exploration of his music and lyrics.
Music Magal
MusicMagal is a group recommendation system that recommends n music tracks to a group of m users taking all of the m users preferences into account. Used last.fm data, an alternating least squares model and item2vec embeddings. After computing and when the resulting playlist is output, we create a real playlist using Spotify Web API.
RECENT EXPERIENCE
Outbrain
Machine Learning Engineer
2019 - present
AutoML team, Recommendations group. Owner of various internal products that help data scientists automate their work, allowing them to deploy models at a large scale. End to end responsibilities: backend and frontend development, machine learning research, MLops & product management.