AI Adversarial Specialist - Fully Remote | Upto $35/hr

Imagine yourself at the forefront of AI innovation, not just building but safeguarding the future of artificial intelligence. Mercor, a dynamic company headquartered in San Francisco and backed by visionary investors like Benchmark, General Catalyst, and Peter Thiel, is actively connecting exceptional creative and technical talent with top-tier AI research labs. We are on the hunt for sharp minds to join us as an AI Adversarial Specialist, a fully remote role where your expertise will directly shape the robustness and safety of cutting-edge AI models worldwide. If you're passionate about probing systems for vulnerabilities and ensuring AI operates ethically and securely, this opportunity is designed for you.

Overview

Mercor is seeking an insightful AI Adversarial Specialist to join our fully remote, global team. This contract position offers a competitive hourly rate and the unique chance to contribute to critical AI safety initiatives for leading research labs. You will be instrumental in identifying and mitigating risks within conversational AI models, ensuring they are resilient against various forms of exploitation. This is a truly worldwide opportunity, inviting talent from any location to collaborate asynchronously and make a significant impact.

Key Responsibilities

In this vital role, you will get to:

Red Team AI Models: Dive deep into conversational AI models and agents, actively seeking out vulnerabilities. Your focus will include jailbreaks, prompt injections, common misuse cases, and sophisticated bias exploitation techniques.
Generate High-Quality Human Data: Create crucial human-generated data. This involves meticulously annotating observed failures, classifying different types of vulnerabilities, and flagging systemic risks for further investigation.
Apply Structured Testing: Work within established taxonomies, benchmarks, and testing playbooks. This ensures a consistent and thorough approach to every adversarial engagement.
Document Reproducible Findings: Produce clear, actionable reports, comprehensive datasets, and detailed attack cases. These outputs enable our customers to understand and implement necessary corrective actions effectively.
Work Independently and Asynchronously: Tackle higher-sensitivity projects with autonomy. You will benefit from clear guidelines and access to wellness resources, supporting your work in a flexible environment.

Requirements

To excel in this position, you should bring:

Fluent Language Skills: Proficiency in both English and Thai is essential for this role.
Prior Red Teaming Experience: Demonstrated experience in AI adversarial work, cybersecurity, or socio-technical probing. Your background should include actively testing and breaking systems.
Strong Communication Skills: The ability to clearly articulate complex risks and findings to diverse audiences is paramount.

What You'll Gain

Joining Mercor means becoming a part of a forward-thinking community where your contributions truly matter. Here's what you can expect:

Impactful Work: Play a direct role in making AI safer and more robust for everyone.
Competitive Compensation: Earn between $24 and $35 per hour for your valuable expertise.
Flexibility: Enjoy the freedom and work/life balance that comes with a fully remote, asynchronous contract position.
Professional Growth: Collaborate with and learn from leading AI research labs, expanding your skill set at the cutting edge of technology.
Global Team: Work with an elite, diverse talent pool from around the world.