Research Scientist, Machine Learning and Eye Tracking

Meta - Redmond, WA

Full-time - Senior

Redmond, WA

Web Search Portals, Libraries, Archives, and Other Information Services

About the position

The Research Scientist in Machine Learning and Eye Tracking at Meta focuses on developing advanced machine learning and computer vision methods for eye tracking in augmented and virtual reality (AR/VR) systems. This role involves working with large-scale datasets to create robust models that cater to diverse user needs, optimizing performance across various sensor modalities, and collaborating with a multidisciplinary team to innovate in the field of eye tracking technology.

Responsibilities

• Research and develop novel machine learning and computer vision based methods for eye tracking in AR/VR form factors.
• Work with large-scale eye tracking data sets to build models that work across the diversity of the entire human population.
• Combine data driven and 3D model based methods to deliver robust performance independent of device fit.
• Incorporate data from multiple sensor modalities to build robust, efficient models.
• Quantify the robustness and accuracy of these models using statistical analysis of the data.
• Optimize models to deliver robust, always-on eye tracking in power-constrained platforms.
• Categorize failures and mine the data for difficult edge cases, and use this to improve the worst-case performance of eye tracking algorithms.
• Help to develop simulation and ground truth data collection tools that reduce the amount of data needed to deliver robust models.
• Collaborate with other research scientists, hardware and software engineers to develop innovative machine learning techniques for eye tracking use-cases.

Requirements

• PhD degree in computer science, machine learning, or equivalent experience in a related field.
• Experience building and shipping CV, ML, or optimization systems that are proven to be robust in real world use cases.
• Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience.
• 5+ years experience in machine learning for computer vision applications.
• 3+ years of programming experience in Python and experience with ML frameworks such as PyTorch.

Nice-to-haves

• Experience with large-scale machine learning techniques like semi-supervised learning, weakly-supervised learning and online adaptation of ML models.
• Experience building personalized models for Eye Tracking or other human tracking tasks such as body, face or hand tracking.
• Experience delivering robust data-driven models incorporating multiple sensor modalities at different data rates.
• Understanding of sensor and optics principles that impact computer vision applications, including MTF, quantum efficiency, refraction, reflectance, etc.
• Experience with transformers for vision applications.
• Proven track record in publishing papers in machine learning and/or computer vision conferences including CVPR, ICCV, ISMAR, ECCV, NeurIPS, AAAI, ICLR.

Benefits

• Health insurance
• Paid holidays
• Flexible scheduling
• Professional development
• Wellness programs