Research Scientist Intern, Multimodal Behavioral Modeling, Hearing Science (PhD)

The Meta Reality Labs Research Team brings together a world-class team of researchers, developers, and engineers to create the future of virtual and augmented reality, which together will become as universal and essential as smartphones and personal computers are today. In particular, the Meta Reality Labs Research Audio team works on technologies that help people hear and communicate better in everyday life, from spatial audio to hearing assistance on AI glasses. Recent work includes Conversation Focus, which helps people hear the person they're talking to in noisy environments. As wearable devices become part of daily life, understanding how people actually listen, attend, and communicate in real-world settings is central to building audio experiences that adapt to each person. We are looking for interns who want to model how people listen and communicate in the real world. Using multimodal, egocentric data from wearable devices (eye and head movement, speech and vocal behavior, gesture, and the acoustic scene), the intern will build models that infer listening and communication states such as auditory attention, listening effort, and conversational engagement. Ground truth for these states is often sparse or ambiguous. The goal is to inform hearing assistance and audio technologies that adapt to the listener. The role suits candidates with a background in machine learning, computer science, hearing science, or auditory or cognitive neuroscience, who want to work at the intersection of those fields. Successful candidates will combine strong quantitative and machine learning skills with an understanding of human hearing and communication. Specific projects will be designed with successful applicants to match their skills and experience to current research needs. Our internships are twelve (12) to twenty-four (24) weeks long and we have various start dates throughout the year.

Responsibilities

• Develop novel machine learning and statistical models that infer listening and communication behavior from multimodal wearable data (for example gaze, head movement, vocal cues, and turn-taking dynamics), with applications to hearing assistance.
• Relate model outputs to established hearing science measures, such as speech intelligibility, listening effort, and self-reported communication difficulty, to validate what the models capture.
• Design data collection strategies, benchmarks, and metrics for models trained on limited or ambiguous ground truth, including data from listeners with and without hearing loss.
• Design, run, and analyze human subjects experiments in laboratory and naturalistic listening environments.
• Collaborate with researchers and engineers across hearing science, audio signal processing, machine learning, and perception science.
• Communicate the research agenda, progress, and full analyses of results.

Minimum Qualifications

• Currently has, or is in the process of obtaining, a PhD in machine learning, computer science, hearing science, or auditory or cognitive neuroscience, or a related field
• 2+ years of experience with Python or a similar language for data analysis and modeling
• 2+ years of experience in statistical modeling and/or machine learning at the doctoral level, including time-series or sequential data
• Research experience with one or more of the following: multimodal behavioral data (e.g., eye tracking, motion, physiological, or speech signals), speech perception, auditory attention, or conversational communication
• Must obtain work authorization in the country of employment at the time of hire, and maintain ongoing work authorization during employment Proven track record of results as demonstrated by grants, fellowships, patents, or first-authored publications at venues such as ARO, ASA/JASA, Ear and Hearing, Trends in Hearing, JSLHR, ICASSP, Interspeech, ICMI, ETRA, CHI, NeurIPS, or similar
• Experience with deep learning frameworks such as PyTorch, including multimodal or sequence models
• Experience with data from wearable or egocentric sensors (e.g., eye tracking, IMU, head-worn microphones), especially in naturalistic or conversational settings
• Experience with hearing-impaired populations, hearing aids, or cochlear implants
• Experience modeling under limited data, weak labels, or ambiguous ground truth
• Experience working with large, high-dimensional datasets from multiple sources
• Experience working and communicating cross-functionally in a team environment
• Intent to return to a degree program after the completion of the internship/co-op