ByteDance
Research Scientist Graduate (Seed Model - Multimodal Interaction & World Model) - 2027 Start
San Jose, California · Full-time · Posted Oct 28, 2026
- Experience
- No years specified.
- Education
- BS in a listed or related discipline; MS is preferred.
About the job
Develop multimodal foundation and world models that support reasoning, planning, perception, memory, decision-making, tool use, and human-system interaction.
What you'll do
- Develop multimodal foundation models integrating vision, language, audio, and environment signals; design and optimize world models for reasoning, planning, and interaction.
- Build training pipelines for data curation, alignment, and reinforcement learning; improve agent capabilities including perception, memory, decision-making, and tool use.
- Explore next-generation interaction paradigms between humans and intelligent systems.
What they ask for
- Completing or recently completed a BS degree in Computer Science, Electrical Engineering, Electrical and Computer Engineering, Physics, Mathematics, or a related discipline.
- Excellent coding ability and knowledge of data structures and fundamental algorithms; proficiency in C/C++ or Python, etc.
- Demonstrated interest or project experience in relevant areas.
Nice to have
- Completing or recently completed a related MS degree.
- Internship experience in multimodal learning, reinforcement learning, or agent systems.
Tools
C/C++, Python
AI focus
Multimodal foundation models, world models, reinforcement learning, and agent capabilities for intelligent interaction.
Model-written summary · the original posting is authoritative