patrick pérez
Scholar

patrick pérez

Google Scholar ID: 8Cph5uQAAAAJ
kyutai
computer visionimage processingmachine learningartificial intelligence
Citations & Impact
All-time
Citations
22,108
 
H-index
58
 
i10-index
140
 
Publications
20
 
Co-authors
116
list available
Publications
20 items
Browse publications on Google Scholar (top-right) ↗
Resume (English only)
Academic Achievements
  • Feb 2025: Kyutai introduced Hibiki, the world’s first on-device, high-fidelity simultaneous speech-to-speech translation model.
  • 2024: Multiple papers accepted at top-tier conferences including NeurIPS’24, ECCV’24, ICML’24, and CVPR’24.
  • Jul 2024: Launched Moshi, the first publicly released real-time voice AI model.
  • Jun 2024: Survey paper on unsupervised object discovery accepted by IJCV.
  • 2023: Papers accepted at ICCV’23, NeurIPS’23, CVPR’23, ICLR’23, etc.
  • 2022: Published work at CoRL’22, ECCV’22, ICIP’22, with code releases for key papers.
  • Frequent keynote speaker, tutorial presenter, and workshop organizer at major conferences like CVPR, ECCV, and ICCV.
Research Experience
  • Dec 2023–present: CEO at Kyutai.
  • 2018–2023: VP of AI and Scientific Director of valeo.ai at Valeo.
  • 2009–2018: Research scientist at Technicolor.
  • 2004–2009 and 1993–2000: Research scientist at Inria.
  • 2000–2004: Research scientist at Microsoft Research Cambridge.