$ cat industry.yaml
-
period:
Fall 2018 - Present
title:
Staff Research Engineer
organization:
Google DeepMind
description:
I’ve been part of Google DeepMind in Zurich since 2018 (originally Google Brain). My focus has shifted over the years: from large-scale computer vision and transfer learning, to sparse Mixture-of-Experts architectures for vision and multimodal models, and now to Gemini pretraining.
-
period:
Summer/Fall 2017
title:
Software Engineer Intern
organization:
Google
description:
I interned at Google Research on the Video Content Analysis team, applying deep learning to content-based video recommendation to help fix the “fresh start” problem for recently uploaded videos in traditional recommendation systems.
-
period:
Summer/Fall 2015
title:
Software Engineer Intern
organization:
Google
description:
Working with the StreetSmart team at Google Geo, we significantly improved the models for business change detection from Street View imagery, based on Convolutional and LSTM neural networks. This allows to keep the listings of businesses in Google Maps fresh at a lower cost.
-
period:
Summer 2014
title:
Computer Vision Engineer
organization:
Blinkfire Analytics
description:
Worked on the backend system for brand logo detection on millions of images from social networks. Designed the system to use multiple classification models, and improved both the speed and the accuracy of the results. Blinkfire Analytics uses computer vision to measure brand impact on social media, providing reports to sport teams, players, agents and sponsors.
-
period:
Summer 2013
title:
Software Engineer Intern
organization:
Google
description:
Working at Google Research with the OCR team, I developed a framework for synthetic training data generation for many OCR applications. We used the framework to improve the training for Google Books, and trained state-of-the-art models for Ads OCR and in-the-wild OCR, using only synthetic data.
-
period:
Summer 2012
title:
Software Engineer Intern
organization:
Google
description:
I joined the OCR team at Google Research to develop one of the first systems for feature extraction, based on deep learning. Thanks to this work, the character error rate was reduced significantly across more than 40 languages for Google Books.
$ cat education.yaml
-
period:
2013 - 2018
title:
Ph.D.
organization:
Universitat Politècnica de València
description:
I did a Ph.D. at the Pattern Recognition and Human Language Technology Research Center. The main result of my PhD was “A Probabilistic Formulation of Keyword Spotting”, but I also worked on general handwritten text recognition.
-
period:
2012 - 2014
title:
Master's degree
organization:
Universitat Politècnica de València
description:
Master in Artificial Intelligence, Pattern Recognition and Digital Imaging working on keyword spotting for historical handwritten text documents. Thesis: “Out of Vocabulary Queries for Graph-based Keyword Spotting”.
-
period:
2007 - 2012
title:
Engineer's degree
organization:
Universitat Politècnica de València
description:
Engineering degree in Computer Science, focusing on Formal Languages and Artificial Intelligence. I spent one year at the Royal Institute of Technology (KTH), in Sweden doing Machine Learning and Distributed Systems courses.