AI Safety Specialist - Remote Position

About Mercor

Mercor bridges top-tier creative and technical experts with premier AI research institutions. Based in San Francisco, the company is backed by notable investors including Benchmark, General Catalyst, Peter Thiel, Adam D'Angelo, Larry Summers, and Jack Dorsey.

Position Overview

This contract role involves remote work with a focus on enhancing conversational AI safety and security.

Key Responsibilities

• Execute red teaming activities on conversational AI models and agents, concentrating on vulnerabilities such as jailbreaks, prompt injections, misuse scenarios, and exploitation of biases.
• Create and annotate high-quality human-generated data identifying model failures, vulnerability classifications, and systemic risk flags.
• Utilize structured approaches through taxonomies, benchmarks, and playbooks to ensure uniformity and rigor in testing procedures.
• Prepare comprehensive and reproducible documentation including reports, datasets, and examples of attacks to enable client actions.

Required Qualifications

• Fluency in both English and Indonesian languages.
• Previous experience in red teaming within AI adversarial contexts, cybersecurity, or socio-technical evaluations.
• Strong communication skills capable of clarifying risks to both technical and non-technical audiences.

Preferred Skills and Experience

• Background knowledge in Adversarial Machine Learning, Cybersecurity, or socio-technical risk assessment.
• Creative probing skills such as those gained from psychology, acting, or writing to approach adversarial thinking unconventionally.

Application Process

• Submit a resume upload.
• Complete an AI-based interview tailored to your resume.
• Fill and submit the application form.

Additional Information

• The interview process takes approximately 20 to 30 minutes.
• Applicants should complete all steps for consideration as applications are reviewed daily.
• For inquiries or support, contact [email protected].