AI Safety Experts – English, Marathi

This is a fully remote position, open to applicants in United States.

📋 Description

• Conduct red-team assessments on conversational AI models and agents utilizing jailbreaks, prompt injections, misuse scenarios, bias exploitation, and multi-turn manipulation.

• Generate human data by annotating failures, classifying vulnerabilities, and identifying systemic risks.

• Implement taxonomies, benchmarks, and playbooks to ensure consistent testing.

• Create reproducible reports, datasets, and attack scenarios for clients.

• Review AI outputs related to sensitive subjects such as bias, misinformation, and harmful behaviors.

• Detect vulnerabilities that automated testing may overlook.

• Increase evaluation coverage and minimize unexpected production issues.

• Assist in training and advancing frontier AI systems for Mercor's AI lab and enterprise clientele.

⛳️ Requirements

• Fluent/native proficiency in English and Marathi.

• Strong discernment regarding language and content.

• Capability to evaluate whether AI responses are accurate, comprehensive, and suitable, along with the ability to articulate the reasoning.

• Meticulous attention to subtle errors, inconsistencies, and gaps.

• Consistent adherence to guidelines and quality standards.

• Ability to clearly communicate reasoning to both technical and non-technical audiences.

• Flexibility across various projects, task types, and clients.

• Status as an independent contractor.

• H1-B and STEM OPT candidates are not eligible.

• Nice-to-have: experience in adversarial machine learning, including jailbreak datasets, prompt injection, RLHF/DPO attacks, and model extraction.

• Nice-to-have: background in cybersecurity, including penetration testing, exploit development, and reverse engineering.

• Nice-to-have: experience with socio-technical risks, such as harassment/disinformation probing, abuse analysis, and conversational AI testing.

• Nice-to-have: experience in psychology, acting, or writing to foster unconventional adversarial thinking.

🏝️ Benefits

• Fully remote position.

• Flexible work schedule allowing you to manage your own time.

• Weekly payments through Stripe or Wise.

• Project durations may vary depending on needs and performance.

• Reasonable accommodations available upon request.

• Opportunity to gain experience in human data-driven AI red teaming.

• Direct involvement in enhancing the robustness, safety, and trustworthiness of AI systems.

• Competitive compensation.

• Collaboration with leading researchers.