Gabriele Berton

Training Vision Language Models

United States

About

I create innovative solutions to real-world challenges, leading me to publish in top conferences while delivering impactful industry applications.

Experience

  • Research Engineer at Google DeepMind
    Mar 2026 - Present · 4 mos

  • Postdoctoral Scientist at Amazon
    Mar 2025 - Mar 2026 · 1 yr 1 mo

  • Applied Scientist at Amazon
    May 2022 - Aug 2022 · 4 mos

    Improving an existing action spotting pipeline in videos through multi-modal neural networks Analyzing video understanding and audio processing methods for a real-world application

  • Research Fellow at Italian Institute of Technology
    Apr 2020 - Jun 2021 · 1 yr 3 mos

    Research of deep learning algorithms for visual geo-localization, which is the task of finding the location where a given photo was taken. Development of a large-scale software to perform visual geo-localization, which integrates state-of-the-art methods from a number of computer vision and machine learning fields, such as image classification, semantic segmentation and efficient nearest neighbor search.

  • Software Engineer at Consoft Sistemi
    Dec 2017 - Dec 2018 · 1 yr 1 mo