强化学习算法工程师-剪映CapCut
Core
Research and develop video creation Agents using Large Language Models (LLM) and Vision-Language Models (VLM), focusing on alignment methods, Agent design, and end-to-end training via Agentic Reinforcement Learning (RL).
Role type
Senior IC Reinforcement Learning Algorithm Engineer (Video Creation Agents)
Builds
Video creation Agents and related AI features for CapCut
Domain
Generative AI, Video Production, Reinforcement Learning
Deliverable
production ML models
Required skills
Reinforcement Learning algorithms, LLM/VLM training, Data mining and cleaning, Reward function design, Transformer architectures, Stable Diffusion, Code implementation, Algorithmic problem solving
Preferred skills
ACM/ICPC/NOI/IOI/Top Coder/Kaggle awards, Publications in NeurIPS/ICML/ICLR/ACL/EMNLP, Experience with Tool-Use/MCP, Creative generation projects
Technologies
RLHF, SFT, DiT, ViT, Transformer, Stable Diffusion
