Meta AR/VR Job | Research Scientist, Perceptual Audio and Machine Learning | Quest
Citys（岗位城市）: Sunnyvale, CA
The perception of audio is at the center of our research, as it informs our work in spatial audio, speech intelligibility, defining audio quality, and many of the AR/VR systems. As a Research Scientist on the Advanced Audio Technologies Team at Facebook Reality Labs, you will be tackling the perceptual and psychoacoustic and ML challenges that define cutting-edge augmented & virtual reality, conducting research at the intersection of basic and applied science, as it informs our technology directions and identifies opportunities for revolutions in our audio quality metrics and audio experiences.
We are seeking a Research Scientist who is well versed in audio machine learning and auditory perception, and is excited about doing research that helps us inform AR/VR audio requirements and metrics towards new technologies.
We are seeking an individual primarily with a hands-on skill set in audio modeling and machine learning in some of the following areas: speech intelligibility, audio perceptual evaluation, design of experiments, computational modeling of the audio system, audio quality metrics, intelligibility, distortion, timbre, dynamics and loudness perception, etc.
Bachelor’s degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience.
PhD in machine learning, computer science, hearing science, acoustics, acoustic engineering, perceptual psychology, psychoacoustics, neuroscience, or a related field with an emphasis on perceptual modeling, audio machine learning psychoacoustics/physics, and/or metrics evaluating perceptual performance
3+ years of experience with machine learning (ML) or AI in audio (PyTorch, TensorFlow, Keras, Scikit-learn, Python, and associated audio processing toolboxes)
3+ years experience with MATLAB, Python, or similar program audio data analysis
Practical experience in setting up and managing a full ML modeling pipeline (Github, etc)
Knowledge of the human auditory system and experience applying this model to novel situations and predict how the ear, brain, and body will respond
Responsible for modeling different aspects of audio quality using a wide range of data science modeling approaches as well as state of the art ML based techniques
Collaborate with colleagues to structure efficient data collection across a wide range of speech and audio applications
Developing ML architectures to model audio datasets (identification, classification and regression models) using CNN, DNN and other modern ML techniques
Work with cross-functional partners in both research and product development to understand and prioritize modeling needs for greatest product impact
Employ existing knowledge of human sensory encoding and audio quality metrics to help guide modeling in the most efficient and meaningful manner
Communicate and share the findings from modeling cross functionality for maximum benefit across the HW organization
Explore opportunities to develop and expand the role of ML based modeling in new technology and product development
Knowledge of perceptual audio evaluation techniques, standard audio quality metrics, and deriving new metrics for audio applications
Experience with statistical analysis tools (e.g. SPSS, SAS, R, XLstat, and/or JMP)
Experience with user studies (psychoacoustics) and human factors
Familiarity with audio system and/or VR, MR, AR platforms
Experience in working on a team that ships product
Experience with research published in journals and international conference presentations or publications