算法策略工程师（大模型方向） - 抖音平台产品
Core
Designing and optimizing algorithms for Douyin's AI news large model, including content generation, retrieval-augmented generation, and automated evaluation systems.
Role type
Senior LLM algorithm strategy engineer
Builds
AI news agents, automated evaluation systems, and domain-adapted large models for news and author assistant scenarios.
Domain
Social media / Large Language Models / Content Generation
Deliverable
production ML models
Required skills
LLM Prompt Engineering, LoRA/QLoRA fine-tuning, RLHF, DPO/GRPO alignment, RAG, Search Recommendation, Agent/Tool Use, SQL, Python, Machine Learning, Deep Learning
Preferred skills
Experience with fact-checking, automated evaluation, complex LLM application deployment, reinforcement learning for optimization
Technologies
LoRA, QLoRA, DPO, GRPO, RLHF, RAG, SQL, Python
Responsibilities
Build reusable AI news agents and automated evaluation systems; Optimize information supply and distribution strategies across multiple scenarios; Improve AI author assistant response quality using reinforcement learning; Develop operational platforms to automate author discovery and strategy execution; Conduct domain adaptation and customized training of large models using fine-tuning and alignment techniques.
Seniority
Senior, hands-on IC
