大模型平台研发（数据工程） - TikTok内容安全平台
Core
Building AI data infrastructure for multimodal large model training, evaluation, and safety governance in international content security scenarios.
Role type
Senior IC AI Data Engineer (Multimodal)
Builds
TB~PB-scale offline and real-time data pipelines, multimodal data assets, and synthetic data for model training and evaluation.
Domain
AI/ML, Content Safety, Big Data
Deliverable
production ML models
Required skills
Python/Go/C++, Data Modeling, ETL, Data Quality, Spark, Flink, Iceberg, Ray, Multimodal Data Processing, Data Synthesis
Preferred skills
Large Language Model principles, SFT/RLHF, Multimodal Alignment, Data Augmentation
Responsibilities
Lead multimodal data modeling, governance, and assetization for AI training and inference; Build metadata, lineage, quality, and cost governance systems; Construct scalable data pipelines for model iteration; Explore data synthesis and augmentation to generate high-quality samples; Collaborate with algorithm and product teams to design data solutions for risk governance.
