AI Adversarial Specialist - Fully Remote

About the Role

Mercor partners elite creative and technical professionals with top AI research organizations. The company is headquartered in San Francisco and backed by prominent investors such as Benchmark, General Catalyst, Peter Thiel, and Jack Dorsey.

This contract position offers a remote working arrangement and focuses on enhancing AI model safety through adversarial testing and evaluation.

Key Responsibilities

• Conduct adversarial testing on conversational AI systems, including jailbreak attempts, prompt injection tactics, and bias identification.
• Produce detailed human-annotated data by marking failure points, categorizing vulnerabilities, and highlighting systemic AI risks.
• Utilize structured tools like taxonomies, benchmark tests, and established playbooks to standardize evaluation procedures.
• Create reproducible documentation consisting of reports, datasets, and attack scenarios to support client decision-making.
• Assess AI-generated content on sensitive issues like bias and misinformation, with possible involvement in projects requiring greater discretion.

Qualifications

• Fluency in both English and Vietnamese is mandatory.
• Prior experience in red teaming, AI adversarial operations, or cybersecurity domains is required.
• Capability to clearly communicate potential risks to audiences with varied technical understanding.
• Preferred skills include knowledge of adversarial machine learning, cybersecurity practices, and socio-technical risk factors.
• Additional advantageous skills encompass creative probing, psychological insight, acting, or writing to support unconventional adversarial approaches.

Application Process

• Submit a resume as part of the initial application.
• Complete an AI-based interview tailored to the submitted resume.
• Fill out and submit the application form to complete the process.

Additional Information

• For interview and platform details, candidates may consult the company’s documentation.
• Support inquiries can be directed to [email protected].
• The recruitment team evaluates applications daily; completing all steps expedites consideration.