Remote Adversarial Prompt Engineer for Safer AI

OpenTrain AI, Inc.; United States; Remote; USD 62,000 - 90,000; Part time

Job summary

OpenTrain AI, Inc. is seeking an Adversarial Prompt Expert to probe large language models for unsafe outputs, jailbreaks, and guardrail weaknesses. This is a remote, part-time role based in the United States with flexible scheduling and hourly pay up to $65.

You will craft adversarial prompts, review model responses, and document findings to support stronger AI defenses and engineering decisions. Prior experience in adversarial testing, red teaming, or security mindset is highly valued.

Qualifications

• Experience testing LLMs for unsafe or undesired behavior.
• Ability to document test plans and results clearly.
• Familiarity with open-source and closed-source language models or similar systems.

Responsibilities

• Probe multiple language models to identify weaknesses.
• Create adversarial prompts and jailbreak attempts across topics.
• Review model responses for failures and guardrail gaps.
• Record attempts and outcomes with clear documentation.
• Research relevant topics and contribute insights for improved defenses.

Skills

Adversarial testing

Red teaming

Security mindset

Prompt engineering

Documentation

Written communication

Persistence

Job description

OpenTrain AI, Inc. is seeking an Adversarial Prompt Expert to probe large language models for unsafe outputs, jailbreaks, and guardrail weaknesses. This is a remote, part-time role based in the United States with flexible scheduling and hourly pay up to $65.

You will craft adversarial prompts, review model responses, and document findings to support stronger AI defenses and engineering decisions. Prior experience in adversarial testing, red teaming, or security mindset is highly valued.