Research Scientist, World Models Graduate (Intelligent Creation) - Global Frontier Tech Recruitment Program - 2027 Start (PhD)
Core
Researching world models and agentic foundation models to enable autonomous, long-horizon multimodal content creation (image/video generation, editing, and understanding) for TikTok users.
Role type
Research Scientist (PhD level)
Builds
Multimodal foundation models, agentic RL systems, and long-context video generation pipelines
Domain
Generative AI, Multimodal AI, World Models, Agentic AI
Deliverable
production ML models
Required skills
Python, C++, algorithms, large-scale model training, post-training (SFT/RL), agentic architecture, tool-calling optimization, temporal consistency modeling, causal relationship modeling
Preferred skills
GenAI, multimodal foundation models, Embodied AI, first-author publications in top-tier venues (CVPR, ICLR, NeurIPS, SIGGRAPH, ICML)
Technologies
Python, C++, SFT, RL, agentic frameworks
Responsibilities
Develop multi-modal data generation pipelines; Build training pipelines for long-context interactive video generation; Advance video generation models for physical dynamics and object interactions; Explore new AI-driven products
Seniority
PhD level, Research Scientist
