Enforcement Detection Analyst, Youtube, Trust and Safety

Google Google · Big Tech · San Bruno, CA +1

This role focuses on developing and implementing strategies to enforce YouTube policies at scale, leveraging Machine Learning and Large Language Models (LLMs) for abuse detection and mitigation. The analyst will collaborate with cross-functional teams, provide incident response, and design prompts for LLMs to improve content classification.

What you'd actually do

  1. Develop and implement strategies and solutions to enforce YouTube policies at scale. Provide data-driven insights and recommendations for effective abuse prevention and enforcement. Leverage Machine Learning, as appropriate.
  2. Collaborate closely with stakeholders and partners in engineering, product and policy to drive complex anti-abuse solutions involving and impacting various cross-functional areas.
  3. Provide rapid incident response for abuse problems, operating on their own. Analyze and draw conclusions on root-causes and define and implement mitigation steps.
  4. Learn complex and technical concepts and systems and deliver meaningful results using them. Communicate technical results and methods clearly.
  5. Design and refine prompts for LLMs to improve their accuracy in identifying and classifying abusive content and behavior. This may include prompt engineering, data labeling, and performance analysis.

Skills

Required

  • data analysis
  • project management
  • Large Language Models (LLMs)
  • spam detection
  • phishing detection
  • platform abuse mitigation

Nice to have

  • machine learning systems
  • problem-solving
  • investigative skills
  • operational improvements
  • generative AI technologies

What the JD emphasized

  • Leverage Machine Learning, as appropriate
  • Experience leveraging Large Language Models (LLMs) to automate the detection and mitigation of platform abuse
  • Proven track record of leveraging LLMs and generative AI technologies to solve business problems

Other signals

  • Leverage Machine Learning, as appropriate
  • Design and refine prompts for LLMs to improve their accuracy in identifying and classifying abusive content and behavior
  • Experience leveraging Large Language Models (LLMs) to automate the detection and mitigation of platform abuse