PeopleofSuperintelligence

Researcher

Aäron van den Oord

Also known as Aaron van den Oord

As of October 2026, Aäron van den Oord is a research scientist at Google DeepMind, known for generative models of audio and images. He completed a doctorate at Ghent University in 2015. In 2016 he co-wrote Pixel Recurrent Neural Networks for ICML and helped introduce WaveNet, a model that generates raw audio one sample at a time. In 2017 he co-wrote Neural Discrete Representation Learning, which introduced the VQ-VAE model, and in 2018 he co-wrote Representation Learning with Contrastive Predictive Coding. The Innovators Under 35 list honored him in 2021, and in December 2024 he co-wrote Google's announcement of new versions of Veo and Imagen.

Photo of Aäron van den Oord

Timeline

  1. 2015

    A doctorate at Ghent University

    He completes a doctorate at Ghent University in 2015, with a dissertation titled Deep architectures for feature extraction and generative modeling.

  2. 2016

    Pixel recurrent neural networks

    He co-writes Pixel Recurrent Neural Networks with Nal Kalchbrenner and Koray Kavukcuoglu for ICML 2016, a paper on modeling the distribution of natural images.

  3. 2016–2017

    WaveNet for raw audio

    In September 2016, he and Sander Dieleman introduce WaveNet, a model that generates raw audio one sample at a time. In November 2017, he, Yazhe Li and Igor Babuschkin describe parallel WaveNet, which is more than 1000 times faster than the original.

  4. 2017

    The VQ-VAE model

    He, Oriol Vinyals and Koray Kavukcuoglu submit Neural Discrete Representation Learning to arXiv in 2017, a paper that introduces the VQ-VAE model.

  5. July 2018

    Contrastive predictive coding

    In July 2018, he, Yazhe Li and Oriol Vinyals submit Representation Learning with Contrastive Predictive Coding to arXiv, a paper that proposes an unsupervised method for learning representations from high-dimensional data.

  6. 2021

    An Innovators Under 35 honor

    The Innovators Under 35 list honors him in 2021, with DeepMind as his organization.

  7. December 2024

    Veo and Imagen updates

    In December 2024, he and Elias Roman co-write a Google blog post that announces new versions of Veo and Imagen.

  8. As of October 2026

    A research scientist at DeepMind

    He is a research scientist at Google DeepMind, a title he also holds in December 2024.

Sources and credits

Photo: avdnoord.github.io; · Aäron van den Oord; creator not stated. License: Not stated by source. Source: avdnoord.github.io. Cropped to a square and resized. Ask us to remove or replace this photo.

  1. Aäron van den Oord - The Mathematics Genealogy Project mathgenealogy.org. Checked 14 October 2026.
  2. Awards | ICML New York City icml.cc. Checked 14 October 2026.
  3. WaveNet: A generative model for raw audio — Google DeepMind deepmind.google. Checked 14 October 2026.
  4. High-fidelity speech synthesis with WaveNet — Google DeepMind deepmind.google. Checked 14 October 2026.
  5. [1711.00937] Neural Discrete Representation Learning arxiv.org. Checked 14 October 2026.
  6. [1807.03748] Representation Learning with Contrastive Predictive Coding arxiv.org. Checked 14 October 2026.
  7. Aäron van den Oord | Innovators Under 35 innovatorsunder35.com. Checked 14 October 2026.
  8. Updates to Veo, Imagen and VideoFX, plus introducing Whisk in Google Labs blog.google. Checked 14 October 2026.