AI Red Teamer (Remote)

We Love AI Jobs; Northern (KY); Remote; USD 67,000 - 198,000; Full time

Job summary

We Love AI Jobs is seeking an AI Red Teamer for a fully remote role based in the USA. You will craft adversarial prompts to stress-test large language models, evaluate guardrails, and document vulnerabilities to strengthen defenses in collaboration with researchers and engineers.

Day-to-day work includes multi-turn prompt design, scoring outputs against harm taxonomies, and contributing to taxonomy development.

Qualifications

• Design adversarial prompts to test AI guardrails across risk categories.
• Evaluate model outputs using structured harm taxonomies and rubrics.
• Collaborate with researchers, engineers, and data scientists on defense improvements.

Responsibilities

• Craft creative prompts and multi-turn scenarios to stress-test AI guardrails.
• Discover ways around safety filters using jailbreak, evasion, and prompt injection techniques.
• Evaluate and score model responses against structured harm taxonomies.
• Document experiments clearly including methods and findings.
• Collaborate with engineers, data scientists, and researchers to strengthen defenses.

Skills

Adversarial prompting

Large language models

Jailbreak techniques

Prompt injection

Model evaluation

Safety rubrics

Python

Structured data annotation

Tools

Python

Job description

# AI Red Teamer (Remote)HandshakeRemote (USA)Full-time$66,560 – $197,600 / yearAI tools:ChatGPTClaudeGeminiUpload your resume once and Scout applies for youApply with ScoutShare this job:LinkedInXSave imageThis remote role stress-tests large language models by creating adversarial prompts and evaluating their responses across safety and risk categories. The work includes documenting vulnerabilities and collaborating with researchers, engineers, and data scientists to improve model defenses.* Contract* Fully Remote### Skills and expertiseAdversarial promptingLarge language modelsJailbreak techniquesPrompt injectionModel evaluationSafety rubricsPythonStructured data annotation### Key responsibilities* Create multi-turn adversarial prompts to test model guardrails across diverse risk categories.* Score model responses against harm taxonomies and document experiments and vulnerabilities.* Collaborate with engineers, data scientists, and researchers to strengthen model defenses.## How this job comparesCompared with 109 open mid-level security jobs on October 10, 2026.Pay: This job pays $67K to $198K a year. The 42 open mid-level security jobs that publish US pay start at $168K on average, so this job starts about 60% lower.Similar open jobs: We Love AI Jobs lists 109 open mid-level security jobs at 65 companies.Remote work: This job is remote. 28 of the 109 mid-level security jobs (26%) are remote.AI in this role: High. AI is the core of the work: building, training or deploying AI systems. The role is centered on directly probing and evaluating large language models every day. It specifically involves using jailbreak, evasion, and prompt injection techniques to identify vulnerabilities.## Full descriptionAI Red Teamer (LLM Generalist)Location: Remote (USA)Employment type: W-2, hourly, employed through TCWGlobalHours: 40 hours per weekAbout the RoleAs an AI Red Teamer, you will stress-test large language models by intentionally trying to break them. Rather than checking whether an answer is correct, you will design creative, adversarial prompts that expose vulnerabilities: unsafe content, bias, broken guardrails, hallucinations, prompt injection weaknesses, and unexpected behaviors. Your work directly supports AI safety and model robustness for leading research labs.This is a generalist red teaming role. You will probe models across the full spectrum of risk categories, including content safety, CBRN (chemical, biological, radiological, nuclear), cybersecurity, persuasion and influence operations, child safety, self-harm, over-companionship, and regulatory compliance. Red teaming may span text, image, voice, and agentic model capabilities depending on project needs.This role requires creativity, curiosity, and an ability to think like an adversary while operating with strong ethical judgment.Day-to-Day Responsibilities• Craft creative prompts and multi-turn scenarios to stress-test AI guardrails across diverse risk categories• Discover ways around safety filters, restrictions, and defenses using jailbreak, evasion, and prompt injection techniques• Explore edge cases to provoke disallowed, harmful, or incorrect outputs• Evaluate and score model responses against structured harm taxonomies and severity rubrics• Document experiments clearly, including what you tried, why you tried it, and what it revealed• Review and refine adversarial prompts generated by other team members• Contribute to harm taxonomy development, calibration exercises, and inter-rater reliability work• Collaborate with engineers, data scientists, and researchers to share findings and strengthen defenses• Work with potentially disturbing content on a regular basis (see Content Warning below)• Stay current on jailbreaks, attack methods, and evolving model behaviorsDesired CapabilitiesCore• Strong hands-on experience using multiple LLMs (ChatGPT, Claude, Gemini, open-source models, etc.)• Intuition for crafting adversarial prompts; familiarity with jailbreak or evasion techniques is a strong plus• Creative, adversarial problem-solving skills• Clear and thoughtful written communication• Strong ethical judgment and the ability to separate adversarial thinking from personal values• Self-directed, collaborative, and comfortable in feedback-heavy environments• Curiosity, persistence, and comfort with frequent failure in experimentationNice to Have• Familiarity with Python or other scripting languages• Experience working with LLM APIs or evaluation tooling• Comfort with structured data annotation and rubric-based scoring• Prior work in trust and safety, content moderation, QA, or security research• Subject matter expertise in any high-risk domain (cybersecurity, chemistry, biology, medicine, law, finance, etc.)You Will Thrive Here If• You treat every model response as a hypothesis to challenge