Position Summary
We are looking for Research Scientists to tackle the data problems that determine our world model capabilities. You will design, develop, and run state-of-the-art systems for curating, ingesting, filtering, annotating, and training with billion-scale visual data. You will work on problems including, but not limited to, motion scoring, camera control, and video understanding. You will translate our team’s human efforts into durable automated systems, whose performance is quantitatively verifiable.
Key Responsibilities
Translate requirements from downstream training, application and eval teams to actionable data work.
Design and build scalable data pipelines.
Collaborate with engineering to scale data pipelines on AWS infrastructure and in-house clusters.
Develop state-of-the-art visual understanding systems to accurately label and annotate images and videos, such as for detecting AI-generated content.
Academic Qualifications
PhD in Machine Learning or Computer Science, or equivalent industry experience.
Professional Experience
Necessary:
Strong communication and collaboration skills for effective cross-functional teamwork.
Ability to navigate ambiguity and drive projects in rapidly evolving research areas.
Exceptional problem-solving and troubleshooting skills to tackle complex technical challenges.
Preferred:
Experience in building and optimizing large-scale video data pipelines.
Experience with large scale video curation.
Experience with data filtering, particularly AI-generated content detection and duplicate detection.
Experience with image or video data annotation.
Experience with video data curriculum.
Strong systems and engineering expertise in deep learning frameworks such as PyTorch.
Research contributions to top-tier conferences or journals (e.g., ICML, ICLR, NeurIPS, ACL, CVPR, COLM, etc.), with published work in relevant domains.