As a Research Scientist in the ASR Research group, the candidate will perform algorithmic research on production-scale datasets with the aim of optimizing accuracy, speed and scalability, primarily for End-to-End (E2E) ASR systems powering Nuance products. This will include research into neural network architectures, training/adaptation algorithms, and data processing / augmentation. Principal duties and responsibilities: Provide experimental and theoretical analysis of speech recognition problems. Formulate/implement new algorithms and design experiments to verify them. Explore new E2E modeling architectures for improving recognition accuracy Optimize the accuracy of E2E ASR systems on challenging real-world data. Follow external developments in the area to keep our research state-of-the-art. Discuss and present ideas and results, reporting progress on a regular basis. Write internal technical reports and papers for external publication. Knowledge, skills and qualifications: Education: MS or PhD in CS or equivalent Minimum years of work experience: 3 for MS or fresh PhD Required skills: Strong background and knowledge of deep learning algorithms. Experience with open source deep learning toolkits (i.e. TensorFlow) Python and shell scripting in a UNIX environment. Good written and oral communications skills. Ability to take initiative, but also follow a plan and work well as part of a team. Preferred skills: Good analytical and diagnostic skills. Experience of running large scale experiments in a grid / cloud GPU computing environment. Experience with sequence to sequence modeling techniques. General understanding to speech recognition algorithms #LI-PS1 #LI-Hybrid Nuance offers a compelling and rewarding work environment. We offer market competitive salaries, bonus, equity, benefits, meaningful growth and development opportunities and a casual yet technically challenging work environment. Join our dynamic, entrepreneurial team and become part of our continuing success.