Visiting Researcher, Methods/Models for Human Perception Responsibilities at Meta (formerly Facebook/Oculus)
Link: https://www.facebookcareers.com/v2/jobs/200816258794061/
Visiting Researcher, Methods/Models for Human Perception Responsibilities
Advance the state-of-the-art in human-in-the-loop experimentation for perception and perceptually-informed outcomes.
Consult and collaborate with partner and client teams to deploy experimentation tooling in real experiments and experiences.
Minimum Qualifications
Currently has, or is in the process of obtaining, a Ph.D. degree in psychology, cognitive science, neuroscience, machine learning, statistics, data science, or a related area, or equivalent work experience.
2+ years research experience in an area with potential applications to human-in-the-loop experimentation in human perception and related domains. Example areas include computational cognitive science, computational neuroscience, psychophysics, active learning, deep Bayesian learning, Gaussian Process models, hierarchical/multitask models, or other relevant areas, as demonstrated via publications (conference or journal), open-source contributions, or similar venues.
1+ years of experience communicating and collaborating with other researchers, as demonstrated by collaborative projects or co-authored research presentations, material (publications/blog posts), or similar.
Must obtain work authorization in the country of employment at the time of hire and maintain ongoing work authorization during employment.
Preferred Qualifications
Experience with crossing disciplinary boundaries, for example from machine learning and statistics to cognitive science, from applied mathematics to neuroscience, or similar, as demonstrated by collaborations across disciplinary domains.
Experience creating research software used by others, for example as part of an academic collaboration, an open-source project, or equivalent.
Experience with SciPy/NumPy and at least one automatic differentiation framework (PyTorch, TensorFlow, JAX, Stan, etc.) -- PyTorch preferred.