Applied Scientist II - Computer Vision
Role details
Job location
Tech stack
Job description
Do you want to join the team working on cutting-edge technology? Text-to-Speech on Device team develops AI-based voice models that run locally on devices. This requires a specific mix of skills between device integration, voice generation technologies, and machine learning.
We are delivering solutions for multiple customers, including offline solutions for Alexa, automotive customers and accessibility voices for visually impaired users. All our models are integrated for devices and work with limited hardware resources.
Responsibilities
Work with the team on end-to-end development of ML models for speech generation, from early experimentation to building production-ready models.
Engage in state-of-the-art and innovative research in areas such as Speech Generation, Gen AI, model compression, and knowledge distillation.
Invent optimization techniques to push the boundaries of deep learning model training and inference.
Create and propose detailed theoretical specifications for novel research ideas and directions, and rigorously justify their correctness.
Train custom Speech Generation and Gen AI models that beat the state-of-the-art and pave the way for developing production models.
Collaborate with other science teams to bring state-of-the-art Speech Generation models from cloud to devices.
About the teamText-to-Speech on Device team is focused on delivering low-footprint AI models for speech generation that can work locally on devices (Android, FireOS, etc.). These models require much less computation power than those hosted in the cloud. We cooperate directly with the teams developing devices and with scientists responsible for the cloud models to provide customers the best possible experience.
Requirements
PhD in engineering, computer science, machine learning, mathematics or equivalent quantitative field
Experience working in Speech Science
Experience applying theoretical models in an applied environment
Experience in state-of-the-art deep learning model architecture design and deep learning training and optimization and model pruning
Experience implementing algorithms using toolkits and self-developed code
Experience with programming languages such as Python, Java, C++
Preferred Qualifications
Experience with model optimization techniques (quantization, distillation, compression, inference optimization etc.)
Experience in professional software development
Experience communicating across technical and non-technical audiences, including executive level stakeholders or clients
Experience working with agile development methodologies
