AI Safety Red Teamer — English & Odia Language Specialist
Paid remote expert bounty in Expert network (Global). Work in the software you already know; rates and ladder published up front.
- $30-$250
- per hr, published up front
- 1
- slots left of 1
- Remote
- no AI experience required
What you'll be paid to show
About the role
We're building a global red team of bilingual language experts to help make AI systems safer, fairer, and more reliable. As an AI Safety Expert fluent in English and Odia, you'll probe large language models with adversarial and edge-case inputs, surface vulnerabilities, and help generate the structured red-team data that our clients use to harden their AI systems against misuse, bias, and harmful outputs.
This is a remote, flexible, project-based role ideal for detail-oriented bilingual professionals who care about the responsible development of AI and enjoy critical, investigative work.
What you'll do
- Craft adversarial prompts and test cases designed to surface unsafe, biased, or misleading AI outputs
- Review and evaluate AI-generated responses in both English and Odia for harmful behaviors, misinformation, and policy violations
- Document findings clearly, categorizing failure modes and severity levels according to project guidelines
- Translate and culturally contextualize prompts and responses to ensure nuance isn't lost across languages
- Collaborate with a distributed team of reviewers to calibrate quality and consistency of red-team data
- Provide structured feedback that directly informs model safety improvements
- Maintain strict confidentiality and follow ethical guidelines when handling sensitive or disturbing content
Requirements
- Native or near-native fluency in both English and Odia (written and spoken)
- Strong analytical and critical thinking skills, with comfort reviewing sensitive or challenging content
- Excellent written communication skills for documenting findings clearly
- Reliable internet access and a quiet remote work setup
- Ability to work independently while following detailed project guidelines
- Prior experience with content moderation, linguistics, translation, journalism, or AI/data annotation is a plus
Compensation
This role pays $20–$22 per hour, with flexible scheduling and remote work from anywhere. Multiple part-time and full-time slots are available for qualified bilingual candidates.
What we're looking for
- AI safety evaluation
- Red teaming
- Bilingual content review (English & Odia)
- Bias and misinformation detection
- Translation and localization
- Adversarial prompt design
- Content moderation
- Data annotation