Remote AI Red Team - Safety & Adversarial Testing

Obsidian; Karachi Division; On-site; PKR 19,466,073 - 27,808,676; Full time

Job summary

Obsidian is seeking a skilled professional to join their red team focused on enhancing AI safety. The role involves red teaming AI models to uncover vulnerabilities, as well as generating human data to drive improvements. Candidates should have prior experience in adversarial work and an instinct for exploring system weaknesses.

Join Obsidian to play a key role in building robust, safe AI systems while gaining valuable experience at the cutting edge of technology.

Qualifications

• Experience in AI adversarial work, cybersecurity, or socio-technical probing is crucial.
• Ability to push systems to breaking points through curiosity.
• Proven communication skills to explain risks to various stakeholders.

Responsibilities

• Conduct adversarial probing of conversational AI models.
• Generate high-quality human data to improve AI systems.
• Follow structured frameworks and document findings clearly.

Skills

Red-teaming experience

Curiosity

Structured methodology

Effective communication

Adaptability

Job description

Obsidian is seeking a skilled professional to join their red team focused on enhancing AI safety. The role involves red teaming AI models to uncover vulnerabilities, as well as generating human data to drive improvements. Candidates should have prior experience in adversarial work and an instinct for exploring system weaknesses.

Join Obsidian to play a key role in building robust, safe AI systems while gaining valuable experience at the cutting edge of technology.