游戏AI-多模态Agent算法专家-UGC场景生成方向
Core
Designing and developing multi-modal Agent frameworks for UGC game scenario generation using frontier AI technologies.
Role type
Senior IC multi-modal Agent algorithm expert (UGC generation)
Builds
Game scenarios and Agent systems
Domain
Gaming + Generative AI
Deliverable
production ML models
Required skills
multi-modal understanding, generative AI, reinforcement learning, Agent framework design, post-training (inference planning, memory management, tool use, multi-turn decision making), PyTorch, SFT, RLHF (PPO, GRPO, DPO), LLM fine-tuning
Preferred skills
LLM Agent system development, top-tier conference publications (NeurIPS, ACL, ICML, EMNLP)
Technologies
PyTorch, LLMs
Responsibilities
Researching multi-modal understanding, generative, and reinforcement learning technologies; Designing and developing Agent frameworks for game scenarios; Enhancing Agent base capabilities including reasoning, planning, memory, and tool calling.
Seniority
Senior, hands-on IC