AI People Wiki

Builders & open source

Heiga Zen

Principal Scientist (from June 2023), Google DeepMind

Heiga Zen is a Principal Scientist at Google DeepMind, a role Zen has held since June 2023, according to Google Research's people page. At Google, Zen was with the Speech team from July 2011 to July 2018 and then joined the Brain team; earlier roles include an intern/co-op researcher position at the IBM T.J. Watson Research Center and a research engineer position at Toshiba Research Europe. Zen was one of the original authors and the first maintainer of the HMM-based speech synthesis system (HTS), and is a co-author of WaveNet: A Generative Model for Raw Audio. Google's page notes that Zen is a fellow of ISCA and IEEE.12

At a glance

Current role
Principal Scientist (from June 2023)
Sector
Company
Work listed
3
Honors
2
Sources cited
7
Profile reviewed

Career

Heiga Zen worked as an intern and co-op researcher at the IBM T.J. Watson Research Center from 2004 to 2005. Zen was then a Research Engineer at Toshiba Research Europe's Cambridge Research Laboratory from 2008 to 2011.1

Zen joined Google's Speech team in July 2011 and joined the Brain team in August 2018. Since June 2023, Zen has been a Principal Scientist at Google DeepMind, with research interests in speech technology and machine learning.1

Research

Zen was one of the original authors and the first maintainer of HTS, the HMM-based speech synthesis system. Zen later co-authored WaveNet (2016), a generative model of raw audio waveforms, and Parallel WaveNet (2017), which made that approach fast enough to be deployed in Google Assistant.123

Zen has also worked on speech datasets and translation. LibriTTS (2019) adapted the LibriSpeech audiobook corpus for text-to-speech research, LibriTTS-R (2023) restored its sound quality, and Translatotron 3 (2023) is an unsupervised approach to direct speech-to-speech translation learned from monolingual speech-text datasets.456

Work and contributions

  • WaveNet Published research

    Co-author of WaveNet: A Generative Model for Raw Audio (2016), with Aäron van den Oord, Sander Dieleman, and others.2

  • HMM-based speech synthesis system (HTS) Open-source code

    Google Research's people page describes Zen as one of the original authors and the first maintainer of HTS. The HTS site says the system's training part is released as patch code to HTK under a free software license.17

  • LibriTTS Open data

    First author of the 2019 paper introducing LibriTTS, a text-to-speech corpus derived from LibriSpeech, and co-author of the 2023 LibriTTS-R paper, which restores the corpus's audio quality.45

Timeline

  1. Began an internship and co-op research position at the IBM T.J. Watson Research Center.1
  2. Joined Toshiba Research Europe's Cambridge Research Laboratory as a Research Engineer.1
  3. Joined the Speech team at Google.1
  4. Co-authored WaveNet, a generative model for raw audio.2
  5. Co-authored Parallel WaveNet on fast, high-fidelity speech synthesis.3
  6. Joined the Google Brain team.1
  7. Released the LibriTTS text-to-speech corpus paper as first author.4
  8. Co-authored Translatotron 3 on speech-to-speech translation with monolingual data.6
  9. Became a Principal Scientist at Google DeepMind.1

Other documented roles

Roles the sources document that are not confirmed as current.

Honors and recognition

  • ISCA Fellow · International Speech Communication Association (ISCA)1
  • IEEE Fellow · Institute of Electrical and Electronics Engineers (IEEE)1

Where they work

Suggest a correction or removal

Profiles cover public professional work only, and each statement cites a source. If something is out of date, missing or wrong, tell us and link a source. People can also ask for their profile to be corrected or removed.

References

7 sources · profile reviewed Oct 21, 2026 · data updated Oct 21, 2026

  1. Heiga Zen — Google Research people pageGoogle Research · Official page · accessed Oct 21, 2026
  2. WaveNet: A Generative Model for Raw Audio (arXiv 1609.03499)arXiv · Paper · published Sep 12, 2016 · accessed Oct 21, 2026
  3. Parallel WaveNet: Fast High-Fidelity Speech Synthesis (arXiv 1711.10433)arXiv · Paper · published Nov 28, 2017 · accessed Oct 21, 2026
  4. LibriTTS: A Corpus Derived from LibriSpeech for Text-to-Speech (arXiv 1904.02882)arXiv · Paper · published Apr 5, 2019 · accessed Oct 21, 2026
  5. LibriTTS-R: A Restored Multi-Speaker Text-to-Speech Corpus (arXiv 2305.18802)arXiv · Paper · published May 30, 2023 · accessed Oct 21, 2026
  6. Translatotron 3: Speech to Speech Translation with Monolingual Data (arXiv 2305.17547)arXiv · Paper · published May 27, 2023 · accessed Oct 21, 2026
  7. HMM/DNN-based Speech Synthesis System (HTS) — project siteHTS working group · Official page · accessed Oct 13, 2026