- Experience
- Any
- Salary
- USD 20 – USD 22 / hour
- Openings
- 1
- Posted
- 4 weeks ago
- Work mode
- Work from home
- Eligibility
- Applicants should be fluent in English and Gujarati and have relevant experience in AI red teaming, cybersecurity, adversarial machine learning, or socio-technical risk analysis. The role is remote and is open to candidates based in Australia.
- Resume
- Required to apply
Job description
About the Role
This opportunity is with Mercor, a company that connects highly skilled creative and technical professionals with top AI research laboratories. The organization is based in San Francisco and is backed by Benchmark, General Catalyst, Peter Thiel, Adam D'Angelo, Larry Summers, and Jack Dorsey.
Position Details
The opening is for AI Safety Experts who can work in English and Gujarati. The engagement is a contract role and can be performed remotely from Australia.
Compensation
The pay range for this role is USD 20 to USD 22 per hour.
What You Will Do
- Test conversational AI systems and agents by attempting jailbreaks, prompt injections, and other misuse scenarios.
- Create reliable human-generated data by labeling failures, identifying weaknesses, and highlighting broader risk patterns.
- Use defined taxonomies, benchmarks, and testing guides to keep evaluations consistent and structured.
- Prepare clear documentation, including reports, datasets, and attack examples that can be used by customers or internal teams.
- Operate independently and on your own schedule while meeting deadlines and contributing to model improvement.
Required Background
- Strong fluency in both English and Gujarati.
- Hands-on experience in red teaming, including AI adversarial work, cybersecurity, or socio-technical probing.
- Ability to communicate risk findings clearly to both technical and non-technical audiences.
Preferred Experience
- Exposure to adversarial machine learning, such as jailbreak datasets, prompt injection, RLHF/DPO attacks, or model extraction.
- Background in cybersecurity areas like penetration testing, exploit development, or reverse engineering.
- Understanding of socio-technical risks such as harassment and misinformation probing, abuse analysis, and conversational AI testing.
Application Steps
The application process usually takes 20 to 30 minutes and includes uploading a resume, completing an AI interview based on the resume, and submitting the form.
Support Information
If applicants need assistance, they can contact support@mercor.com. Additional information about the interview process and platform is available in the provided documentation.
Important Note
The review team evaluates applications daily, so completing every step of the application process is necessary to be considered for the role.