Multi-Modality Research Intern
Core
Pioneer research in multimodal understanding, post-training, reasoning, and agent planning for gaming scenarios.
Role type
Research Intern (Multimodal AI)
Builds
Multimodal large models (VLMs/MLLMs) and research prototypes for in-game applications.
Domain
Gaming / Multimodal AI
Deliverable
research
Required skills
multimodal understanding, vision-language reasoning, LLM/VLM pre- and post-training, reinforcement learning, LLM-based agents, representation learning, experimental design, PyTorch, JAX
Preferred skills
abstracting business scenarios to research problems, deploying prototypes at scale, first-author publications at top-tier conferences
Technologies
PyTorch, JAX
Responsibilities
Pioneer new research directions in multimodal understanding and agent planning; Abstract research problems from game business scenarios; Advance understanding capabilities of multimodal large models; Lead post-training methodologies; Design and execute end-to-end experiments; Collaborate with engineers to deploy prototypes; Publish at top-tier conferences
Seniority
Intern
