后台研发工程师（LBS/POI平台） - TikTok生活服务
Core
Building an online streaming data scheduling system and POI data processing pipeline for TikTok Local Services, integrating Large Language Model (LLM) capabilities for data deduplication, attribute extraction, and content moderation.
Role type
Senior Backend Engineer (LLM Engineering & Data Infrastructure)
Builds
High-throughput POI data ingestion, cleaning, and processing systems; LLM inference services and application infrastructure for local services.
Domain
Internet / Location-Based Services (LBS) / Large Language Models (LLM)
Deliverable
production ML models
Required skills
High-concurrency system design, streaming data scheduling, big data components, LLM engineering (inference, RAG, prompt engineering), multi-modal understanding, system observability, cost optimization.
Preferred skills
POI data processing experience, Agent frameworks, LLM-driven data cleaning/annotation.
Technologies
LLM APIs, Vector search, RAG, Message queues, Databases, Multi-modal models (OCR, Video/Image).
Responsibilities
Design and build online/offline POI data cleaning modules; Implement LLM inference services and governance (Prompt, RAG, structured output); Establish evaluation and cost infrastructure for LLM applications; Collaborate on model deployment and system stability.
Seniority
Senior, hands-on IC