AI基础设施研发工程师-抖音AI
Core
Build high-concurrency, low-latency training/inference/evaluation environments for TikTok's AI agents and Agentic RL systems.
Role type
Senior IC AI Infrastructure Engineer (RL & Agent Systems)
Builds
Scalable training/inference environments, multi-agent collaboration frameworks, and optimized RL infrastructure for TikTok's AI products.
Domain
AI Infrastructure, Reinforcement Learning, Large Language Models
Deliverable
production ML models
Required skills
Java/C++/Python/Go, distributed system design, model optimization, resource scheduling, heterogeneous computing
Preferred skills
YARN/K8S/Ray, LoRA/P-Tuning/RLHF, inference engine optimization, operator optimization, algorithm-engine co-design
Responsibilities
Construct high-concurrency training/inference environments, develop multi-agent collaboration capabilities, optimize training efficiency and memory usage, research and implement cutting-edge AI technologies.
Seniority
Senior, hands-on IC