Roberto Pieraccini

The short version

Speech recognition and spoken language human-computer communication expert, artistic photographer, writer. Born in Italy, lived in Viareggio, Torino, New Jersey, New York State, shortly in Paris, Manhattan, and now in Berkeley, California. Moved from corporate research (CSELT, Bell Laboratories, IBM Research), to startups (SpeechWorks, SpeechCycle), to academic research management (ICSI, Berkeley). Led the conversational team at Jibo. Now Director of Engineering at Google in Zurich, Switzerland.

The longer version

I was born in Italy, grew up in Viareggio, a popular Tuscan beach resort on the Tyrrhenian sea, and earned a degree in Electrical Engineering from the  University of Pisa.  After graduation I moved to the northern industrial city of past royal magnificence known as Torino  to work, as a researcher, at CSELT (Centro Studi E Laboratori Telecomunicazioni), one of the most prestigious telecommunication research labs in Italy at that time.  There I worked on Speech Recognition, in other words building computers that understand speech, and that became my career. It was the 1980s. My son Dan was born in 1985 in a town north of  Torino called Cirié.


Towards the end of the 1980s I felt I wanted to experience living in the US, at least for a while. Because of  interesting research I had done at CSELT, I was invited to spend one year at Bell Laboratories  in Murray Hill, NJ, one among the most prestigious research centers in the world. There  I worked with Larry Rabiner, who is considered to be one of the fathers of the modern speech recognition technology. After a short return to Italy at the end of my year long stay,  Larry offered me a position as a regular member of the technical staff. So I moved to the US for good. That was 1990. My daughter Alex was born in 1991 in Plainfield, New Jersey.  

At  the legendary Bell Labs I worked in a place that had probably one of the highest concentration of  the most brilliant and renowned  scientists in the word in many disciplines.  To name just a few: Dennis Ritchie, the inventor of the C language, Brian Kernighan and Ken Thompson, the inventors of the Unix operating system, Béla Julesz, one of the main contributors to the autostereograms, Arno Penzias, who received the Nobel Prize in 1978 for having discovered the Cosmic Microwave Background Radiation, know also as the 3 Kelvin degree radiation, which validates the Big Bang theory, and many others. 

My major scientific contribution at Bell Laboratories was  that of introducing  statistical models to build systems that could understand the meaning of speech (and not just the words),  and show that in certain task that would work better than any other non-statistical model.  

As a result of the 1996 AT&T's trivestiture, AT&T Labs were created in Florham Park, NJ. I decided to work there. That was 1996. At AT&T Labs I contributed to a new way to look at spoken human-computer communication, that of building machines that learn how to interact by experience by using reinforcement learning theories.

After so many years in corporate research, at the end of the 1990s I started to be attracted by the world of startups. I joined  SpeechWorks as Director of Dialog Technologies. SpeechWorks was then  a small company headquartered in Boston,  a spinoff of MIT, rival to an equally small company called Nuance located in the Silicon Valley. That was 1999. For a while my office was in downtown Manhattan, a few blocks from the World Trade Center. I enjoyed  a spectacular view of the Statue of Liberty and Fridays @4 with Martinis.  I also learned a lot about the commercial world of speech technology.

 I left SpeechWorks before it was acquired by ScanSoft, and joined IBM T.J.Watson Research in YorkTown Heights, NY. That was August 2003. I spent two years at IBM Research where I made great friends and enjoyed living in Westchester county. I worked in multi-lingual search, NLP, and speech recognition. I was the manager of a research group called Advanced Conversational Technologies  but I felt it was time to go back to the startup world. 


Back to Manhattan, back to a startup company which I had followed as an advisor since its inception in 2001. That was SpeechCycle, and I became their Chief Technology Officer. That was August 2005. It was a ride that lasted for more than 6 years.  At SpeechCycle I led innovation in the field of complex dialog systems and I tried to put into practice  some of the things I have been working on as a researcher during my previous lives, but this time with real systems, real customers, real data, and lots of adrenaline.  In 2011 I felt my mission was somehow accomplished, and decided to move on to some other adventure. SpeechCycle was eventually acquired by Synchronoss in 2012. 

In January 2012 I moved West to Berkeley, California, where I became the director and CEO of the International Computer Science Institute, or ICSI, an independent organization for advanced research in many branches of computer science, such as Networking, Security, Speech, Multimedia, Vision, Artificial Intelligence, Computational Biology, Brain Networks, Algorithms  and Computer Architectures. I left ICSI in October 2013. 

In March 2012 my book, The Voice in the Machine, Building Computers That Understand Speech, was published by MIT press. The book, for a general scientifically interested audience, narrates the history and the business of the technology that brought us Siri and Google Voice Search. 

In February 2014 I joined Jibo, Inc., where I lead the team that made the Jibo robot speak and understand speech. I left Jibo in December 2017, and joined Google in Zurich as a Director of Engineering for the Google Assistant team.