Former Researchers

Since the foundation of the Institute in 1991, hundreds of researchers have contributed to its development. We are glad to be here today thanks to their work and support.

We have listed below senior researchers who previously led a research group since Idiap's beginning.



During his 3 years at Idiap, David has led the Machine Learning research group and has been part of numerous PhD theses committee. Since 2007, he has become a fellow of the Alan Turing Institute and professor of machine learning at University College London (UCL). In January 2020, he founded OptimalAI, a company which automate AI decision-making for industry.


During 8 years, Samy was the head of the Machine Learning research group at Idiap. He directed several PhD theses and collaborated on many publications such as the "Discriminative Kernel-based Model to Rank Images from Text Queries". In 2007, he joined the Google research team. Currently, he is working at Apple.


From 1996 to 2023, Hervé was Idiap's director and the head of the Speech & Audio Processing research group. At EPFL, he was Full Professor and headed the LIDIAP laboratory. He also worked as Artificial Intelligence Professor and Scientific Head of Programme for Idiap's master's in Artificial Intelligence in partnership with Swiss Distance University.


During a couple of years, from 1992 to 1994, Thomas developed a research program in computer vision and handwriting recognition at Idiap. Then he worked for IBM, Google and was full professor during ten years at the University of Kaiserslautern in Germany. Since 2016, he is distinguished research scientist at NVIDIA in California.

Barbara CAPUTO

With a PhD in computer science from the KTH Royal Institute of Technology, Barbara came to Idiap to develop her research activities in the field of visual cognitive sciences applied in particular to robotic vision. After Idiap, she continued her career as associate professor and head of the Visual and multimodal applied learning laboratory at the University of Rome "La Sapienza".


From 1992 to 1997, Gérard worked as senior scientist at Idiap, before moving to CNRS in France. He is currently working for various private companies active in technologies development.


After a PhD at EPFL, from to 1991 to 1997, Jean-Luc worked at Idiap as a researcher and as director ad interim during two years. He continued his career at Swisscom and various private companies. Since 2013, he is the head ot the IT unit at the Swiss federal archives.


Ronan did his PhD in machine learning at Idiap and then led the Applied Machine Learning research group. During his time at Idiap, he collaborated and published numerous major papers with a particular emphasis on generic machine-learning. In 2014, Ronan has continued his way as a research scientist at Facebook. Currently, he is working at Apple


After a PhD in computer science at the Queensland University of Technology, John came to Idiap as a research scientist to work on very low rate speech coding, large vocabulary speech recognition and speech synthesis. Since 2014, he started his own business as a consultant in rich transcription and speech recognition and moved later to the Expedia Group where he is currently principal data scientist.


During six years at Idiap, until 1997, Emile was active in the biology field. He is currently president at Bioveyda in California and is focusing on biodiversity issues.

François FLEURET

Francois came to Idiap as senior research scientist to lead the Machine Learning research group. He supervised several doctoral students and collaborated on several major projects such as the MASH European project. Since 2020, François joined the University of Geneva as full professor of computer science.


Idiap welcomed Mike in 2003 to join the Multimodal Interaction research group. During his 6 years at the institute, he notably worked on new techniques for multimodal navigators. Since 2009, Mike moved to Manchester where he founded two companies active in the field of artificial intelligence.


Focusing his research on Gaussian random field modelling and adaptive design of experiments, David was the head of the Uncertainty Quantification and Optimal Design research group at Idiap. During 5 years, he followed several students and took part in the Master in Artificial Intelligence developed by our institute. Since 2020, he is leading a research group at the university of Bern as a titular professor.


Hynek joined Idiap in 2003 to lead the Speech Processing research group. As a senior research scientist, he collaborated on many publications in scientific reviews such as the IEEE Signal Processing Letters or the Journal of Acoustical Society of America. In 2008, he became professor and director at the Center for Language and Speech Processing at The Johns Hopkins University.


During four years, until 1999, Eddy led the machine learning research group at Idiap. He conducted research projects on different pattern recognition and classification techniques. After holding various positions at Motorola, Yahoo and Quantcast, since 2010, he works at Google.


Iain joined Idiap as a senior research scientist in Speech Processing. He participated in several large-scale projects, including IM2, M4 and AMI. His research interests include in particular acoustic antennas and speaker segmentation. In parallel with his research, since January 2004, he has been the scientific coordinator for the integrated European project AMI. He is also an adjunct professor at the University of Queensland and co-founder of Notiv.


Named world leader in robotics research by Scientific American magazine in 2004, José joined Idiap in 2002 to lead research in brain interfaces. During his 6 years here, he directed several PhD theses and collaborated on numerous publications on the theme of brain-computer interfaces. Since 2008 he has been a professor in several universities including EPFL and today holds the post of Carol Cockrell Curran Endowed Chair at The University of Texas at Austin.


From 1998 to 1999, Chafik conducted research at Idiap. He is currently professor of computer engineering at the University of Balamand in Lebanon.


Starting in 1998, for four years, Andrew conducted research in the robust automatic speech recognition field. The techniques included hidden-Markov models, artificial neural networks, multi-stream combination, Bayesian recognition with missing and uncertain data and Bayesian networks. After Idiap he pursued a successful career in both private companies and academia.


Daniel OSHERSON, former director.
After a PhD at the University of Pennsylvania, Daniel managed Idiap from 1991 to 1995 as director. He the continued his career as professor at the University of Pennsylvania and later in Princeton, as a professor of psychology.


Andrei joined Idiap as head of the Natural Language Processing group to continue his researchers on text analysis at the semantic and pragmatic levels for machine translation and information retrieval. He has supervised numerous doctoral theses and received several awards for his publications. Since 2017, Andrei holds the position of professor in computer science at the School of management and engineering Vaud (HEIG-VD).


From 1991 to 1997, Indu research fields at Idiap concerned characterized and integrated liquid crystal light valves performance in optical signal processor machines for real-time signal processing of digital images. After several experiences in Californian private firms focusing on optics, he is currently the principal scientist at Innoveyda in Los Angeles.


During two years at Idiap, from 1992 to 1994, Kari’s research in automatic speech recognition and neural networks. After a career as an adjunct professor at the Arizona state university, since 2008, he is senior principal research scientist at Amazon.


Fabio came to Idiap to work on speech processing. He was part of a PhD thesis committee and participated in many European projects such as RODI (Role based speaker diarization) or RECOD 2012 (Very low bit-rate speech coding). Fabio continued his career as an IT consultant at Libero professionista ICT.

Jithendra VEPA

Jithendra joined Idiap to work in the Speech Processing research group on speech recognition using data-driven approaches, posterior-based hierarchical automatic speech recognition, and speech recognition in meetings. Since 2007, he has worked successively for Philips and Samsung and is now chief scientist at Observe.Ai.


Alessandro joined Idiap as a PhD student and became researcher in the Information Retrieval research group after having obtained his PhD. After his work at Idiap, he became professor at the University of Glasgow and currently occupies in parallel the position of director and principal investigator at Social AI CDT.


Pierre came to Idiap to work in the Multimodal Interaction research group. During 5 years, he developed meeting recording browsers from research in speech, computer vision, and machine learning. He also defined and managed international research projects at multiple universities and corporations, such as AMI & IM2. After he left, Pierre continued as co-founder and CTO of Spiderphone and is now research manager at INFICON.