Skip to content
oruk
Start free

Company

We're building machines that understand people

oruk is a speech lab. We build audio models that transcribe words and classify emotion and speaking style from the same signal. The goal is useful speech understanding with measured, inspectable outputs.

The lab also supports the oruk organization, our nonprofit arm bringing speech technology to endangered languages.

The team

  • Nathan Roll

    Nathan Roll

    Founder & CEO · Research

    Stanford UniversityUniversity of Cambridge
  • Vianney Grenez

    Vianney Grenez

    Co-founder & CTO · Infra

    UC BerkeleyArts et Métiers (ENSAM)
  • Irene Yi

    Irene Yi

    Research

    Stanford UniversityUC Berkeley
  • Büşra Marşan

    Büşra Marşan

    Research

    Stanford UniversityBoğaziçi University
  • Daniel D'Angelo

    Daniel D'Angelo

    Product

    Ernst & YoungCERNUC Santa Barbara
  • Anders Hedberg

    Anders Hedberg

    Growth

    MetaUC Santa Barbara
  • Dillon Silzer

    Dillon Silzer

    Infra

    Microsoft Azure
  • Calbert Graham

    Calbert Graham

    Advisor

    University of CambridgeUC Berkeley
  • Jonas Hedberg

    Jonas Hedberg

    Advisor

    Wells FargoLiftlab
What we believe
  • 01

    Meaning over transcription

    A transcript preserves words but drops acoustic information. oruk models the transcript, emotion, and speaking style together so applications can use both.

  • 02

    Research in the open

    We benchmark against the field on its own terms and publish where it counts. Progress you can verify, not claims you have to take on faith.

  • 03

    Built for builders

    Understanding should be one API call away. We carry the hard parts so the teams shipping conversational AI to real people don’t have to.