Speech Recognition Scientist

Reference Number: 183

Position

 

The role will quantify speech recognition accuracy for many of the currently enabled features in the product and conduct root-cause analysis of accuracy short falls using Nuance analysis tools. You will also be looked upon to recommend a range of techniques for enhancing speech recognition accuracy from major algorithmic enhancements to minor parameter tuning which are backed by quantifiable metrics.
  • Creating recognition test bed
  • Designing experiments
  • Hypothesis testing and analyzing results
  • Summarizing key findings and Making succinct recommendations
  • Assist in recruiting and managing independent pool of outside transcribers
  • Build a speech recognition test laboratory and develop needed tools/scripts to efficiently accomplish results and ensure repeatability
  • Occasionally conduct TOI's with other members of the engineering/Professional Services organizations on key insights

Experience


  • MS or PhD in the field of Speech Recognition
  • Thorough understanding of principles and concepts underlying speech recognition technology
  • Practical insights into the specific Nuance recognition engine used inside the server
  • Intimate familiarity with Nuance analysis tool sets, Grammar Specification Language, recognition framework for Nuance
  • Meticulous fact finder
  • Scientific methodologist
  • Unbiased and factual reporting of results
  • Ability to analyze complex cause-and-effect situations
  • Excellent verbal and written communication skills esp. for technical material
  • 10+ years of very relevant experience in the field of automatic speech recognition and performance tuning
  • 5+ years of experience working specifically on Nuance recognition systems and speech applications