Remote AI Safety Red Team Engineer

Obsidian; Karachi Division; Remote; PKR 22,271,715 - 33,407,572; Full time

Job summary

Obsidian is seeking a dedicated red team member to enhance AI safety by probing models for vulnerabilities. In this role, you will identify biases, document findings, and deliver reproducible artifacts to strengthen AI systems. Ideal candidates are experienced in red-teaming, curious, and skilled in communicating complex risks.

The position is remote, emphasizing collaboration on AI safety as we explore cutting-edge methodologies and serve our clients efficiently.

Qualifications

• Prior red-teaming experience in AI adversarial work or cybersecurity.
• Ability to communicate risks clearly to stakeholders.
• Experience with frameworks or benchmarks in testing.

Responsibilities

• Red team conversational AI models against biases and misuse.
• Generate high-quality human data for vulnerabilities.
• Document testing processes and findings for clients.

Skills

Red-teaming experience

Curiosity and adaptability

Structured approach

Communication skills

Job description

Obsidian is seeking a dedicated red team member to enhance AI safety by probing models for vulnerabilities. In this role, you will identify biases, document findings, and deliver reproducible artifacts to strengthen AI systems. Ideal candidates are experienced in red-teaming, curious, and skilled in communicating complex risks.

The position is remote, emphasizing collaboration on AI safety as we explore cutting-edge methodologies and serve our clients efficiently.