Carlos Bentes

cbentes_pic.png

Hello there!

I’m an AI generalist based in Estonia, with over a decade of experience building ML systems across diverse domains. Most recently I’ve focused on speech AI models for real-time translation, neural codec architectures, and multimodal LLM reasoning.

Feel free to connect with me on one of the social links below and start a conversation.

latest posts

Mar 22, 2026 A-Maze
Feb 23, 2026 JEPA for audio encoding

selected publications

  1. Pinch-AST: Robust Cascaded Speech Translation System for the IWSLT 2026 Simultaneous Speech Translation Task
    Carlos Bentes and Christian Safka
    In Proceedings of the 23rd International Conference on Spoken Language Translation (IWSLT 2026), 2026
  2. Speech Translation Benchmark
    Carlos Bentes
    Jun 2026
  3. JEPA-v0: a self-supervised audio encoder for real-time speech translation
    Carlos Bentes
    Feb 2026
  4. Ship classification in TerraSAR-X images with convolutional neural networks
    Carlos Bentes, Domenico Velotto, and Björn Tings
    IEEE Journal of Oceanic Engineering, Feb 2017
  5. Ship-iceberg discrimination with convolutional neural networks in high resolution SAR images
    Carlos Bentes, Anja Frost, Domenico Velotto, and 1 more author
    In Proceedings of EUSAR 2016: 11th European conference on synthetic aperture radar, Feb 2016
  6. Target classification in oceanographic SAR images with deep neural networks: Architecture and initial results
    Carlos Bentes, Domenico Velotto, and Susanne Lehner
    In 2015 IEEE international geoscience and remote sensing symposium (IGARSS), Feb 2015