AI Safety Expert for English and Thai Red Teaming

Join a red team of human data experts who actively probe conversational AI systems to find weaknesses before they reach users. You will design and execute adversarial text attacks, surface vulnerabilities related to bias, misinformation, and harmful behaviors, and produce reproducible artifacts that help customers harden their models. All work is text based, and higher-sensitivity work is optional and supported with clear guidance and wellness resources.

• Red team conversational AI models and agents, including jailbreaks, prompt injection, misuse scenarios, bias exploitation, and multi-turn manipulation.
• Generate high-quality human data by annotating model failures, classifying vulnerabilities, and flagging systemic risks.
• Follow and apply established taxonomies, benchmarks, and playbooks to ensure consistent testing and coverage.
• Document findings reproducibly, producing reports, datasets, and attack cases that customers can act on.
• Communicate discovered risks clearly to both technical and non-technical stakeholders.

• Prior red teaming experience, such as AI adversarial work, cybersecurity, or socio-technical probing.
• Demonstrated ability to use structured frameworks or benchmarks when testing, rather than ad hoc approaches.
• Strong communication skills, able to explain risks and reproduce attack cases for varied audiences.
• Curiosity and an adversarial mindset, with a habit of pushing systems toward failure modes.
• Adaptability, with the ability to move across different projects and customer contexts.
• Nice-to-have specialties: adversarial machine learning (jailbreak datasets, prompt injection, RLHF/DPO attacks, model extraction), cybersecurity skills (penetration testing, exploit development, reverse engineering), socio-technical risk analysis (harassment, disinformation, abuse in conversational AI), and creative probing skills (acting, psychology, creative writing for adversarial thinking).

• Location: Remote.
• Employment type: hourly engagement.
• All work is text based.
• Work may involve reviewing outputs that touch on sensitive topics such as bias, misinformation, or harmful behaviors.
• Participation in higher-sensitivity projects is optional. Higher-sensitivity assignments are accompanied by explicit guidelines and wellness resources.
• Topics that could be sensitive will be clearly communicated to you before any exposure to that content.

Hourly pay: 24 - 35 hourly.

• Native fluency in English and Thai is required for this position.

• You consistently uncover vulnerabilities that automated tests miss.
• You deliver reproducible artifacts that customers use to improve model safety.
• Evaluation coverage expands and fewer unexpected behaviors reach production as a result of your work.

This role offers hands-on experience in human-driven AI red teaming at the frontier of safety, with a direct impact on making AI systems more robust and trustworthy.