AI Safety Expert (English & Bengali)

Role overview

This contract position is for an AI safety specialist who will evaluate conversational AI systems and agents in English and Bengali. The work is fully remote and focused on identifying weak points, unsafe behaviors, and other reliability issues through adversarial testing.

Pay is set at $20 to $22 per hour, with a weekly commitment that can range from 10 to 40 hours.

Key responsibilities

• Probe chat-based AI models and agents for weaknesses by attempting jailbreaks, prompt injection attacks, and bias-related exploit scenarios.
• Create useful human-labeled data by reviewing failure cases, tagging vulnerabilities, and highlighting broader risk patterns.
• Work with defined taxonomies, benchmarks, and testing guides so evaluations remain consistent and repeatable.
• Record results in a clear, reproducible format by preparing reports, datasets, and attack examples that can be used by customers.

Required background

• Previous experience in AI red teaming, adversarial testing, cybersecurity, or socio-technical analysis is required.
• An investigative, challenge-seeking mindset is important, along with persistence in pushing systems to their limits.
• The ability to work in a methodical, framework-driven way is expected rather than relying on ad hoc approaches.
• Strong written and verbal communication skills are needed to explain technical and non-technical risks clearly.
• Flexibility is important, since the work may shift across different projects and clients.

Application process

The selection process includes three steps: submitting a resume, completing an interview, and filling out a form.

Additional information

This role is a contract remote opportunity in the United States. The position is for candidates who can work in English and Bengali.