AI Red Teamer (Seattle)

Written by the T&S Network from Handshake's posting.

AI red teaming on large language models: adversarial stress-testing of model guardrails across content safety, CBRN, cybersecurity, persuasion, child safety, self-harm, over-companionship and regulatory compliance risks. Covers text, image, voice and agentic capabilities depending on project needs, for leading research labs.

• Self harm and suicide
• Hate speech
• Adult and sexual content
• AI misuse and jailbreaks

What you'd do

• Build inventive prompts and multi-turn conversations that probe guardrails across many risk types.
• Use jailbreak, evasion, and prompt injection techniques to get around safety filters and defenses.
• Push models at the margins to elicit banned, dangerous, or wrong outputs.
• Grade model answers using set harm categories and severity scales.
• Document experiments, including what was tried, why, and what it revealed.
• Review and refine adversarial prompts written by other team members.
• Help build out harm categories, run calibration sessions, and check that raters score consistently.

What it asks for

• Strong hands-on experience using multiple LLMs, including open-source models.
• Intuition for adversarial prompt crafting.
• Clear written communication.
• Familiarity with jailbreak or evasion techniques (preferred).
• Python or other scripting languages (preferred).
• Experience with LLM APIs or evaluation tooling (preferred).

Worth knowing

• Regular and deliberate exposure to harmful content, including violence, self-harm, hate speech, sexually explicit material, and child safety scenarios.
• Support resources are available for this exposure.
• W-2 hourly contract employment through TCWGlobal, 40 hours per week.

Terms in this role

Red teaming.

Testing a system, often an AI model, by deliberately trying to make it fail or misbehave, so problems are found and fixed first.

Jailbreak.

A prompt written to get an AI model to ignore its safety rules.

Our best summary of the posting, and it can be wrong. The employer's posting is what counts: read it in full on Handshakenorth_east before you apply. If we got something wrong, report this listing.