Adversarial Prompt Expert

Overview

You will be part of a red teaming project focused on probing large language models for failure modes and harmful outputs. In this role as Adversarial Prompt Expert, you will craft prompts and scenarios to test model guardrails, explore creative methods to bypass restrictions, and document outcomes systematically.

About Handshake AI

Handshake AI offers a flexible, remote and asynchronous work environment. You will work independently from anywhere with flexible hours and minimal required commitment.

Role Details

• Develop domain-specific prompts and evaluate LLM responses.
• Research topics using AI and learn new skills.
• Collaborate with engineers and safety researchers to improve defenses.
• Document all attempts and outcomes clearly.
• Placement in projects depends on availability.

Qualifications

• Heavy LLM usage with multiple models (open and closed source).
• Expertise in prompt engineering and jailbreaking techniques.
• Adversarial mindset with red teaming or offensive security experience.
• Persistence, creativity and systematic documentation skills.
• Strong ethical awareness and responsibility handling sensitive content.

Work Authorization

This position is open to U.S.-based students, candidates, and recent graduates with US work authorization. F-1 students on CPT/OPT may be eligible, subject to school requirements.

Key skills/competency

• Red teaming
• LLM
• Prompt engineering
• Jailbreaking
• Adversarial testing
• Security
• Research
• Documentation
• Creativity
• Ethical awareness