PeopleofSuperintelligence

Researcher

Joseph Keshet

Also known as Yossi Keshet

Joseph Keshet is a speech researcher whose work spans speech recognition, adversarial examples and speech steganography. He studied electrical engineering at Tel Aviv University, led a speech-processing unit in IDF Unit 8200, and completed a computer science PhD at Hebrew University. His career took him from IDIAP and TTIC to Bar-Ilan and Technion; he also held roles at Amazon, Disruptive AI and aiOla. He coauthored Houdini, Hide and Speak, and Drax, and coedited a book on large-margin speech and speaker recognition. As of October 2026, he is an associate professor of electrical and computer engineering at Technion.

Born
1973
Badges

Years link to their sources.

Timeline

  1. 1991–2008

    Degrees and Unit 8200

    He begins electrical engineering studies in Tel Aviv University's Department of Electrical Engineering-Systems in 1991 and earns a BSc in 1994 and an MSc in 2002. From 1995 to 2002, he heads the speech processing research section at the R&D Center of IDF Unit 8200. He completes a computer science PhD in the Hebrew University's School of Computer Science and Engineering in 2008 under Yoram Singer.

  2. 2007–2012

    Research at IDIAP and TTIC

    In 2007, he becomes a research scientist in speech recognition at IDIAP Research Institute in Martigny, Switzerland. From 2009 to 2012, he is a research assistant professor at TTIC and the University of Chicago's Department of Computer Science.

  3. 2009

    A book on speech recognition

    With Samy Bengio, he coedits Automatic Speech and Speaker Recognition: Large Margin and Kernel Methods, a 2009 book about large-margin and kernel methods for speech and speaker recognition.

  4. 2013–2024

    Bar-Ilan and speech AI roles

    In 2013, he joins Bar-Ilan University's Department of Computer Science as a senior lecturer and heads its Speech, Language, and Deep Learning Lab. In 2020, he becomes a founding venture partner at Disruptive AI and an Amazon Scholar in Tel Aviv. In 2022, he joins the Technion's Faculty of Electrical and Computer Engineering. By April 2024, he is aiOla's Chief Scientist; its AdaKWS keyword-spotting model reaches 95% accuracy, compared with 88% for OpenAI's Whisper.

  5. 2017–2019

    Houdini and hidden speech

    In 2017, he coauthors Houdini, a method for adversarial examples tailored to a task's final performance measure and applied to speech recognition, pose estimation, and semantic segmentation. In 2019, he coauthors Hide and Speak, which uses deep neural networks to hide messages in speech data.

  6. 2025

    Drax parallel speech decoding

    In 2025, he coauthors Drax, a discrete flow matching framework for automatic speech recognition that enables efficient parallel decoding.

  7. 2026

    ISCA Distinguished Lecturer

    In 2026, the International Speech Communication Association names him a Distinguished Lecturer.

  8. As of October 2026

    Associate professor at Technion

    As of October 2026, he is an associate professor of electrical and computer engineering at Technion. He also serves as Deputy Editor for Speech Processing at IEEE Transactions on Audio, Speech, and Language Processing.

Career

  1. Since 2022TechnionAssociate Professor of Electrical and Computer Engineering
  2. Since 2007Idiap Research InstituteResearch Scientist in Speech Recognition
  3. From 2020Disruptive AIFounding Venture Partner
  4. From 2020AmazonAmazon Scholar
  5. 2013–2022Bar-Ilan UniversitySenior Lecturer
  6. 2009–2012Toyota Technological Institute at ChicagoResearch Assistant Professor
  7. 2009–2012University of ChicagoResearch Assistant Professor
  8. 1995–2002Israel Defense ForcesHead of the speech processing research section
  9. aiOlaChief Scientist

Education

  1. 2008Hebrew University of JerusalemPhD in computer science
  2. 2002Tel Aviv UniversityMSc
  3. 1994Tel Aviv UniversityBSc

From the biography.

  1. 16% of AI career moves in the 2020s went from a university to a companyIncludes Joseph Keshet. We estimate that 16% of AI career moves in the 2020s went from a university to a company; 29% of moves starting at a university went to a company, and 76% ended the university role.
  2. Universities keep AI researchers for a median of 6 years, compared with 2 years at frontier labsIncludes Technion. Across 15,811 dated work stays (6,529 people), median role tenure runs from 6 years at universities to 2 years at frontier labs. We compare all five organization types, starting decades from the 1970s to the 2020s, and the organizations that keep people longest.
  3. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Technion. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  4. 41% of people who worked abroad later worked in their country of birth among people with a recorded birth country and dated work historyIncludes Technion. Of 991 people with a recorded birth country who worked outside it, 410 (41%) later held a work position at home, led by the United States at 69%.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Joseph Keshet yet. Send us a free-licensed photo.

  1. Joseph Keshet en.wikipedia.org. Checked 21 October 2026.
  2. Automatic Speech and Speaker Recognition: Large Margin and Kernel Methods | Wiley uat.store.wiley.com. Checked 21 October 2026.
  3. TTIC Faculty Alumni ttic.edu. Checked 21 October 2026.
  4. Yossi Keshet - Israeli Research Community Portal cris.iucc.ac.il. Checked 21 October 2026.
  5. aiOla's Speech AI Technology Outperforms OpenAI's Whisper in Recognizing Jargon prnewswire.com. Checked 21 October 2026.
  6. Houdini: Fooling Deep Structured Prediction Models arxiv.org. Checked 21 October 2026.
  7. Hide and Speak: Towards Deep Neural Networks for Speech Steganography arxiv.org. Checked 21 October 2026.
  8. Drax: Speech Recognition with Discrete Flow Matching arxiv.org. Checked 21 October 2026.
  9. ISCA Distinguished Lecturers 2026 isca-speech.org. Checked 21 October 2026.
  10. Joseph Keshet - Technion - Israel Institute of Technology cris.technion.ac.il. Checked 21 October 2026.
  11. IEEE Transactions on Audio, Speech, and Language Processing Editorial Board signalprocessingsociety.org. Checked 21 October 2026.