About the Role
The role involves challenging AI models to uncover harmful outputs, biases, or security flaws, particularly in Indonesian-language contexts, to improve system robustness and safety.
Responsibilities
- Identify potential weaknesses in AI-generated responses using Indonesian language inputs
- Design and execute test scenarios that simulate malicious or unintended user behavior
- Analyze model outputs for safety violations, misinformation, or cultural insensitivity
- Provide structured feedback to improve AI alignment and reduce risks
- Collaborate with linguists and engineers to refine evaluation frameworks
- Document findings with clear examples and severity assessments
- Contribute to the development of red teaming methodologies for multilingual AI systems
- Stay current with AI safety research and emerging threat patterns
- Test models across diverse domains including social interaction, content generation, and decision support
- Ensure compliance with ethical guidelines during adversarial testing
- Report critical vulnerabilities through secure internal channels
- Assist in validating fixes and monitoring reoccurrence of issues
- Work across time zones to support global team coordination
- Use scripting or automation tools to scale testing efforts where appropriate
- Maintain confidentiality of sensitive model behaviors and internal processes
Compensation
Competitive salary and benefits package
Work Arrangement
Full-time, remote position with flexible hours
Team
Part of a global AI safety and evaluation team
Language Requirements
- Fluency in Indonesian is required, including familiarity with regional dialects and informal speech patterns
- Professional working proficiency in English for collaboration and reporting
Security Clearance
- Candidates must pass a background check
- Handling of sensitive AI behaviors requires strict confidentiality
Available for qualified candidates
