Software Engineer Iii, Genai Safety

Google Google · Big Tech · Zürich, Switzerland

Software Engineer III, GenAI Safety at Google DeepMind Abuse team. Focuses on designing, building, and scaling account-centric detection systems within the GenAI Safety Platform to protect Generative AI products from misuse. Develops classifiers for policy violations and abusive behavior, integrates with enforcement mechanisms, and analyzes data to improve detection methods. Requires ML/AI experience and software development skills.

What you'd actually do

  1. Design, build, and maintain components of our account-centric abuse detection systems on the GenAI Safety Platform.
  2. Develop and implement classifiers to identify policy-violation and abusive behavior in near real-time, leveraging account-level signals.
  3. Collaborate with other engineers to integrate detection systems with enforcement mechanisms like strikes and limited disabilities.
  4. Analyze data to understand abuse patterns and improve the effectiveness of our detection methods.
  5. Write and test production-quality code, and participate in code reviews.

Skills

Required

  • Software development in Python, C++, Go or Java
  • Machine Learning or Artificial Intelligence

Nice to have

  • large-scale data processing pipelines and tools (e.g., Flume, SQL)
  • anti-abuse systems, fraud detection, or security
  • AI driven development and Agentic workflows

What the JD emphasized

  • mission-critical team
  • intersection of Generative AI and abuse fighting
  • shape the safety and trustworthiness of groundbreaking AI products
  • develop innovative solutions to novel abuse vectors specific to GenAI
  • ensure AI is used responsibly

Other signals

  • GenAI Safety Platform
  • detecting misuse of Generative AI products
  • account-centric detection systems
  • novel abuse vectors specific to GenAI