AIGC后训练算法专家-智能创作
Core
Research and optimize post-training algorithms for AIGC models to enhance text-to-image, video, and 3D generation capabilities for core business applications.
Role type
Senior IC AIGC post-training algorithm expert
Builds
Vertical large models for image/video generation, 3D creation, and intelligent editing tools
Domain
Generative AI, Computer Vision, Multimodal Learning
Deliverable
production ML models
Required skills
Diffusion models, VLM PEFT, RL algorithms (GRPO, PPO, DPO, RM), Post-training algorithms, Data construction for large models, Multimodal understanding
Preferred skills
Visual large model research and deployment, Text/visual/multimodal technical directions, Product landing experience
Technologies
SFT, RL, VLM-Based PE systems
Responsibilities
Explore multimodal understanding, generative, and reinforcement learning technologies; Develop SFT and RL techniques for text-to-image/video/3D generation; Optimize vertical large models using data synthesis and preference alignment; Build VLM-Based PE systems to support core business model deployment.
