- Role: Red Team Operator (English & Chinese) - Remote
- Location: 100% Remote (Global)
- RATE: USD 50.50/ HOUR
Why This Role Exists
One of our clients is building a dedicated red team to proactively identify weaknesses in advanced AI systems. The safest AI systems are those that have already been rigorously stress-tested by skilled human adversaries.
This role focuses on adversarial testing of conversational AI models, identifying vulnerabilities before they surface in production environments. You will generate structured red-team datasets that directly strengthen AI robustness and safety.
Some projects may involve reviewing AI outputs related to sensitive topics (e.g., bias, misinformation, harmful behaviors). All work is text-based. Participation in higher-sensitivity tracks is optional, clearly communicated in advance, and supported by structured guidelines and wellness resources.
What You'll Do
- Conduct adversarial testing of conversational AI systems (jailbreaks, prompt injections, misuse cases, bias exploitation, multi-turn manipulation)
- Generate high-quality red-team data by annotating failures and classifying vulnerabilities
- Identify systemic weaknesses and recurring risk patterns
- Apply structured taxonomies, benchmarks, and testing frameworks
- Produce clear, reproducible documentation including reports, datasets, and attack cases
- Deliver actionable findings for AI research and safety teams
Who You Are
- Experienced in AI red teaming, adversarial testing, cybersecurity, or socio-technical risk analysis
- Naturally adversarial and curious you instinctively push systems to edge cases
- Structured in your approach you rely on benchmarks and repeatable methodologies
- Strong communicator able to explain complex risks to both technical and non-technical stakeholders
- Adaptable and comfortable working across multiple projects
Nice-to-Have Specialties
- Adversarial ML: jailbreak datasets, prompt injection attacks, RLHF/DPO vulnerabilities, model extraction
- Cybersecurity: penetration testing, exploit development, reverse engineering
- Socio-technical risk: harassment probing, misinformation testing, abuse analysis
- Creative adversarial thinking: psychology, narrative construction, unconventional probing strategies
What Success Looks Like
- You uncover vulnerabilities automated testing pipelines miss
- You produce reproducible artifacts that measurably improve AI safety
- Evaluation coverage expands across linguistic and cultural edge cases
- AI systems deploy with fewer safety blind spots
Engagement Details
- Rate: $50.50/hour (aligned with expertise and project sensitivity)
- Fully remote and asynchronous
- Independent contractor engagement
- Projects may be extended, shortened, or concluded based on performance and research needs
- Weekly payments via Stripe or Wise
- Visa sponsorship (H1-B or STEM OPT) not available
All qualified applicants will be considered without regard to legally protected characteristics. Reasonable accommodations are available upon request.
APPLY NOW!
