多模态内容理解算法工程师-Global Monetization Product and Technology
Core
Develop foundational multi-modal content understanding algorithms for commercial scenarios, providing semantic embeddings and text labels for ad delivery, creative optimization, and content moderation.
Role type
Senior IC multi-modal algorithm engineer (commercialization)
Builds
High-quality semantic embeddings and text label systems for ad delivery and content moderation
Domain
Advertising technology and multi-modal AI
Deliverable
production ML models
Required skills
Python, C++, PyTorch, LLMs, MLLMs, Diffusion/Flow-Matching models, Reinforcement Learning, Multi-modal RAG, Visual Chain-of-Thought, Multi-modal Agents, Reward Models, Video/Image understanding, Cross-modal retrieval and alignment, Multi-modal model training and fine-tuning
Preferred skills
Experience in ad/ecommerce/recommendation scenarios, International conference publications, GitHub projects, Kaggle/ACM awards
Technologies
PyTorch, LLMs, MLLMs, Diffusion models, Flow-Matching, RL, RAG, CoT, Agents
Responsibilities
Develop multi-modal content understanding algorithms for ad delivery and content moderation; Explore advanced multi-modal capabilities like RAG and Agents for commercial tasks; Drive collaborative evolution of understanding and generative models; Track and implement frontier technologies like Deep Research and Multi-Agent systems.
