Idiap has a new opening for a Postdoc position in Deep Neural Network applied to Speech Recognition

The Idiap Research Institute seeks qualified candidates for a postdoc position in the field of Deep Learning, Deep Neural Networks, and Compressive Sensing, applied to speech processing and automatic speech recognition.

Directly reporting to Prof. Hervé Bourlard, the research will be conducted in the framework of a recently granted Swiss NSF project, in collaboration with 2 PhD students. The research will focus on novel speech recognition and speech modeling approaches using novel deep neural network and sparse recovery (compressive sensing) models, towards a principled, generalized, hierarchical model ranging from feature extraction to lexical and syntactic modeling (possibly also exploiting word embedding approaches). While being funded by Swiss NSF, the present project will also take place in related EU and US (IARPA) projects.

The Postdoc candidate should hold a PhD in relevant fields, while demonstrating excellence and motivation in research in general, and in the targeted research areas. He should all have (at least) a background in statistics or applied mathematics, optimization, linear algebra and signal processing. Postdoc applicants should ideally have good expertise in hidden Markov models (HMMs), Deep Neural Networks (DNNs), and/or Sparse Recovery Modeling. The applicants should also have strong programming skills and be familiar with MATLAB, C/C++, and various scripting languages such as Python and with the Linux environment.

The postdoc contract will be for one year, with a start salary of 80,000. - CHF/year, with a reasonable chance to be renewed based on performance.

Position will remain open until it has been filled, ideally before the end of 2018. All queries related to the advertised position can be sent to Prof. Hervé Bourlard (bourlard at idiap.ch).

Interested? Please submit your candidature through the Idiap online recruitment system: Postdoc position in Deep Neural Network applied to Speech Recognition