Platform Safety Data Scientist – Risk & Insights

TikTok USDS Joint Venture; San Jose (CA); On-site; USD 144,000 - 329,000; Full time

Job summary

TikTok USDS Joint Venture LLC is seeking a data scientist to design and optimize metrics and detection pipelines for platform safety. You will apply statistical methods to assess security policies and inform risk decisions, while collaborating with product, operations, and algorithm teams to turn insights into concrete safety improvements.

You will also help build reusable data analysis tools, productize analytics capabilities, and explore AI/LLM-driven enhancements to workflows.

Qualifications

• 2+ years of data science experience in Internet/content platform safety
• Proficient in SQL and Python, with strong data visualization skills
• Familiar with causal inference and A/B testing; strong communication & cross-team collaboration

Responsibilities

• Design, build, and maintain core metrics systems and data pipelines for platform safety and risk.
• Use rigorous statistical methods to evaluate security policies and inform decisions.
• Develop quantitative analysis systems and translate insights into actionable improvements for safety.
• Create reusable analysis frameworks and contribute to data product design.
• Leverage AI & LLM techniques to improve operational workflows.

Skills

SQL

Python

Data visualization

A/B testing

Causal inference

Education

Master's degree or above in statistics, data science, computer science, operational research, mathematics, or related field

Tools

Pandas

NumPy

Jupyter

Job description

TikTok USDS Joint Venture LLC is seeking a data scientist to design and optimize metrics and detection pipelines for platform safety. You will apply statistical methods to assess security policies and inform risk decisions, while collaborating with product, operations, and algorithm teams to turn insights into concrete safety improvements.

You will also help build reusable data analysis tools, productize analytics capabilities, and explore AI/LLM-driven enhancements to workflows.