Amazon is looking for an applied scientist for the Paralinguistics Applied Science team with a strong background in Deep Learning, Automation Speech Recognition or Natural Language Processing.
A day in the life
As an Applied Scientist in the team, you will work on state of the art deep learning models that understand disfluent or dysarthric speech and help modulate Alexa's voice so that its work for everyone. We take pride in publishing our work in leading speech conferences like Interspeech and ICASSP and help improve the state of the art of the speech industry. The ideal candidate will use their background in speech recognition to build Robust ASR models as well as classification models to detect disfluent speech. They will work in large scale distributed system that works on millions of hours of speech data, while respecting the privacy of our users. We welcome scientists from other backgrounds like Computer Vision, Natural Language Processing, Reinforcement learning and Operations Research to apply for this role. In this role, you'll get an opportunity to advance your skills in ASR, Reinforcement learning, Bayesian Statistics and Classification models.
About the hiring group
The Paralinguistics Applied Science team helps Alexa understand "How" our customers are speaking and helps modulate it voice accordingly. We build models that enhance ASR engine to be robust to dysarthria or disfluencies in speech. Our North Star is to make Alexa instantly familiar, ever-present personal assistant, advisor, and companion who works equally well for everyone. Alexa understands everyone seamlessly through natural interactions that accommodates the users and can offer additional assistance for customers with speech impairment, such as speech therapy and being the voice of communication for customers in need. We are the team, that works on giving Alexa a human-like, Lombard Effect, where it adjusts its volume accordingly to the environment.
· Develop deep learning models to help Alexa understand disfluent speech
· Analyze key trends through billions of datapoints across languages and regions
· Deploy production grade models to production and respond/incorporate customer feedback
Amazon is committed to a diverse and inclusive workplace. Amazon is an equal opportunity employer and does not discriminate on the basis of race, national origin, gender, gender identity, sexual orientation, protected veteran status, disability, age, or other legally protected status. For individuals with disabilities who would like to request an accommodation, please visit https://www.amazon.jobs/en/disability/us.
· PhD or equivalent Master's Degree plus 4+ years of experience in CS, CE, ML or related field
· 2+ years of experience of building machine learning models for business application
· Experience programming in Java, C++, Python or related language
· Master's degree in Computer Science, Electrical Engineering, or related technical field
· 3+ years industry experience applying Machine Learning techniques (SVM, GMM, LDA, etc.) to solve real problems
· 1+ years experience with deep learning technology
· Knowledge of data structures, algorithm, and information retrieval; programming proficiency (e.g. Python) is required
· Track records of publishing papers and patents -- the applicant should at least has one first-author paper in tier-1 conference.
· PhD in Electrical Engineering, Computer Sciences, or related technical field
· Industry experience in speech, speaker recognition or image recognition.
· Experience defining system architectures and exploring technical feasibility trade-offs
· Academic and/or industry experience with standard ML techniques, training pipelines, Neural Net frameworks
· Strong verbal/written communication, including the ability to effectively communicate with both business and technical teams