微信搜索-后台开发工程师-表情搜索
Core
Develop core engineering modules for the WeChat emoji search online retrieval system, including LLM planner, generative retrieval, personalized long-sequence, and RankMixer modules.
Role type
Senior backend engineer (LLM-powered search)
Builds
High-availability distributed search infrastructure for WeChat emoji retrieval
Domain
Social media search, Large Language Models (LLM)
Deliverable
production ML models
Required skills
C/C++, data structures and algorithms, LLM inference optimization, distributed system architecture, network communication optimization, VLLM
Preferred skills
Experience with massive-scale service development, performance optimization for LLM/VLLM inference chains
Responsibilities
Develop core modules for online retrieval and offline processes (content ingestion, feature extraction, indexing), design high-availability distributed architectures, collaborate across frontend, algorithm, and operations teams