BEGIN:VCALENDAR
VERSION:2.0
PRODID:Icfo
X-PUBLISHED-TTL:P1W
BEGIN:VEVENT
UID:69d1248526741
DTSTART:20250527T130000Z
SEQUENCE:0
TRANSP:OPAQUE
DTEND:20250527T140000Z
LOCATION:Seminar Room
SUMMARY:ICFO | CARLOS FLORENSA
CLASS:PUBLIC
DESCRIPTION:ABSTRACT:\nCarlos Florensa returns to ICFO to share the arc of 
 his career&mdash\;from Summer Fellow in science communication and biomedic
 al optics to multimodal AI research at OpenAI. In this talk\, he will give
  a high-level\, candid perspective on his path from ICFO to a PhD at UC Be
 rkeley\, his five years at the robotics startup Covariant (acquired by Ama
 zon)\, and his current work on real-time video-speech-to-speech intelligen
 ce.\nCarlos will also discuss broader trends in artificial intelligence&md
 ash\;what's already possible\, what's coming next\, and how this rapid evo
 lution could impact basic sciences\, innovation\, and entrepreneurship.\nB
 IO:\nCarlos Florensa started his academic journey at the UPC studying Indu
 strial Engineering and Mathematics\, and first joined ICFO in 2012 as a Su
 mmer Fellow in the KTT department\, creating a science-based magic show to
  explain ICFO research &mdash\;from CosmoCaixa to London. He returned in 2
 015 as a Severo Ochoa Fellow\, working with Turgut Durduran on computer vi
 sion for non-contact pulse-oximetry.\nHe went on to pursue a PhD in Artifi
 cial Intelligence at UC Berkeley\, with internships at DeepMind and NVIDIA
 . He then joined the robotics startup Covariant\, where he built AI system
 s until its acquisition by Amazon. Today\, Carlos leads video speech-to-sp
 eech research at OpenAI\, working on the frontier of fluid real-time intel
 ligence.
DTSTAMP:20260404T144733Z
END:VEVENT
END:VCALENDAR