Vision-Language Robotics Research Engineer

XDOF; San Mateo (CA); On-site; USD 140,000 - 210,000; Full time

Benefits offered by this job

Competitive compensation

Equity in growth

Comprehensive health benefits

Flexible work arrangements

Collaborative team environment

Job summary

XDOF in San Mateo is seeking a Research Engineer/Scientist to drive data-centric robotics research at the intersection of vision-language models and robot learning. You will build end-to-end systems that convert egocentric and teleoperation video into high-signal training data for VLMs, and contribute to models themselves.

You will explore metadata discovery, structured annotations, and curricula to improve cross-embodiment transfer and world models, then translate insights into scalable

Qualifications

• MS or PhD in Computer Science, Robotics, Machine Learning, or a related field from a top-tier program.
• 3–7 years of research or applied research experience in vision-language models, video understanding, robot learning, or related areas.
• Deep fluency in PyTorch and experience with large-scale training infrastructure.

Responsibilities

• Design and implement vision-language pipelines for egocentric and teleoperation video, including captioning, grounding, and semantic annotation at scale.
• Develop representations bridging visual perception, language, and robot action across VLAs, video prediction, and world models.
• Build data curation systems to assess quality, diversity, and coverage of large-scale robot demonstration datasets.
• Work with hands-on manipulation data, real teleoperation footage, and sim-generated rollouts.
• Collaborate with partner labs to define data requirements and improve downstream policy performance.
• Stay current on VLAs, video foundation models, and related architectures, translating insights into production systems.

Skills

PyTorch

Software engineering fundamentals

Education

MS/PhD in CS/Robotics/ML

Tools

PyTorch

Job description

XDOF in San Mateo is seeking a Research Engineer/Scientist to drive data-centric robotics research at the intersection of vision-language models and robot learning. You will build end-to-end systems that convert egocentric and teleoperation video into high-signal training data for VLMs, and contribute to models themselves.

You will explore metadata discovery, structured annotations, and curricula to improve cross-embodiment transfer and world models, then translate insights into scalable