- 经验
- 任何
- 薪水
- 每小时 20 至 22 美元
- 职位空缺
- 1
- 发布
- 2周前
- 工作模式
- 在家办公
- 合格
- Candidates with experience or interest in AI adversarial testing, cybersecurity, or socio-technical probing who can work remotely and are able to operate in English and Punjabi.
- 恢复
- 需要申请
职位描述
Role overview
This is a remote contract opportunity for AI Safety Experts fluent in English and Punjabi. The work focuses on evaluating conversational AI systems and agents for safety issues, with pay set at $20 to $22 per hour. The expected weekly commitment ranges from 10 to 40 hours.
What you will do
- Probe conversational AI models and agents for weaknesses by testing jailbreak paths, prompt injection attacks, and bias-related failures.
- Create labeled human data by reviewing failures, classifying different types of vulnerabilities, and identifying broader systemic risks.
- Use defined taxonomies, benchmarks, and testing playbooks to keep evaluations consistent and repeatable.
- Capture results in a clear, reproducible format through reports, datasets, and concrete attack examples that can be used by customers.
What we are looking for
- Hands-on background in AI red teaming, adversarial testing, cybersecurity, or socio-technical analysis.
- A naturally investigative and adversarial approach that helps uncover edge cases and failure modes.
- A disciplined, framework-based testing style rather than ad hoc experimentation.
- Ability to communicate risk findings effectively to both technical teams and non-technical audiences.
- Comfort working across different projects and client contexts with flexibility and adaptability.
Application process
- Submit a resume.
- Participate in an interview.
- Complete the application form.
Additional information
This role is a contract position with remote work arrangements. Compensation is hourly at $20 to $22. The expected workload is 10 to 40 hours per week. The language requirement specified for the position is English and Punjabi.