AI Data Lead (World Models & Multimodal AI)

Location: Shenzhen, China (On-site)

About the Role

We're building next-generation World Models capable of perception, reasoning, interaction and long-horizon simulation.

As our AI Data Lead, you will own the complete multimodal data lifecycle—from strategy and acquisition to governance, evaluation and continuous improvement—building the data engine that powers our models.

You will work closely with research teams across World Models, Agents and Multimodal Foundation Models.

Responsibilities

• Define the overall data strategy for large-scale World Model and multimodal model training.
• Build and continuously improve high-quality datasets across images, videos, text, 3D assets, simulation environments and interaction trajectories.
• Design scalable data collection, cleaning, annotation and quality assurance pipelines.
• Develop benchmark datasets and evaluation protocols that directly support model iteration.
• Lead data governance, metadata standards, quality metrics and lifecycle management.
• Partner closely with Research Scientists and ML Engineers to translate research needs into scalable data solutions.
• Manage external data vendors and strategic data partnerships globally.
• Optimize data production efficiency, infrastructure cost and resource utilization.
• Establish internal standards for data security, compliance and confidential asset management.

Requirements

• 5+ years of experience in AI data, machine learning data engineering or multimodal data systems.
• Proven experience building training datasets for LLMs, Vision-Language Models, World Models or Agent systems.
• Deep understanding of data collection, annotation, curation, filtering and quality control pipelines.
• Experience designing evaluation datasets and benchmark methodologies.
• Familiar with multimodal data including image, video, text, audio, 3D or simulation data.
• Strong cross-functional communication with research and engineering teams.
• Experience leading data teams or managing complex cross-functional AI data projects.

Preferred Qualifications

• Experience at frontier AI labs or foundation model organizations.
• Experience supporting World Models, embodied AI, robotics or simulation systems.
• Familiarity with synthetic data generation and automated data pipelines.
• Python scripting and data automation experience.
• Experience with global data vendors and international collaboration.
• Passion for building scalable AI data infrastructure from the ground up.

Why Join Us

• Build the data engine behind next-generation World Models.
• Work alongside researchers pushing the frontier of multimodal AI and autonomous agents.
• Flexible work environment with an engineering-first culture.
• Competitive compensation and long-term equity opportunities.
• Access to large-scale compute resources and ambitious research projects.