Jobs › Topics › Red teaming and safety
AI red teaming and AI safety jobs
21 open roles name red teaming, adversarial testing, AI safety or content safety work today, on 6 platforms. The median listed hourly pay is $70, from 17 roles that list an hourly rate or range. We counted on . See how we count.

- Open roles
- 21on 6 platforms
- Median listed pay
- $70per hour, across 17 roles
- $100 an hour or more
- 8roles that list it
Who these roles are for
The board lists 21 open roles across Alignerr, Mercor and micro1 and other platforms. Security testers, writers, researchers, and bilingual reviewers often look for this work. Many roles ask for careful reading, clear writing, and consistent judgment. Some roles ask for language skills, research experience, or technical testing experience. Read the posting's testing scope and requirements before applying.
Open roles that match this guide
Open on . Pay as each listing states it.
micro1 · 58 countries
AI Jailbreak & Prompt-Injection Security Expert
$50–$90/hrNew to micro1? Create an account before applying.
Mercor · 107 countries
LLM Research Scientist (Pre-training & Computer Vision & Adversarial Robustness)
$100–$120/hrNew to Mercor? Create an account before applying.
Turing
AI Benchmark & Red-Teaming Specialist
Pay not listedMercor
Bilingual Chinese Generalist Expert — AI Safety
$48–$52/hrNew to Mercor? Create an account before applying.
Mercor
Bilingual Spanish Generalist Expert — AI Safety
$43–$47/hrNew to Mercor? Create an account before applying.
Platforms with these roles
Each row is a platform with such a role open today, the one with the most first. The median is of listed hourly pay, from 8 or more roles that state it.
What red teaming work looks like
Red teaming means testing AI models for unsafe answers within each platform's rules.
- Many roles involve reviewing model answers against the safety criteria in a posting.
- Some roles involve writing test prompts and recording how a model responds.
- Many roles involve labeling harmful content and explaining the concern in clear language.
- Some roles involve checking answers across languages or subject areas.
- Some postings describe disturbing content that the work shows.
Kinds of roles
Red teaming and adversarial testing: the title names red teaming, jailbreaks, adversarial or penetration testing. AI safety and policy: the title names AI safety, safety review or policy work. A role counts once, in the first kind its title fits.
| Kind | Roles | Platforms | Median |
|---|---|---|---|
| Red teaming and adversarial testing | 14 | 5 | $90 |
| AI safety and policy | 7 | 4 | – |
What the roles list for pay
The median listed hourly pay is $70, based on 17 roles with an hourly rate or range. The median is the middle figure, and a listed range counts at its midpoint. The middle half of listed hourly pay runs from $45 to $90. The highest listed hourly figure is $120. Each platform sets its own pay and can change it.
Roles open today
In the board's Recommended order. Each title opens the role's page, with its full posting and its Apply link.
- AI Safety & Evaluation Specialist$30–$50/hrmicro1 · Cybersecurity · US
- AI Jailbreak & Prompt-Injection Security Expert$50–$90/hrmicro1 · Cybersecurity · 58 countries
- Behavioral Health Expert, AI Safety and Model Evaluation$45–$70/hrMercor · Science & Research · US
- LLM Research Scientist (Pre-training & Computer Vision & Adversarial Robustness)$100–$120/hrMercor · Data Analysis · 107 countries
- AI Benchmark & Red-Teaming SpecialistPay not listedTuring · Other
- Bilingual Chinese Generalist Expert — AI Safety$48–$52/hrMercor · Language & Audio
- Bilingual Spanish Generalist Expert — AI Safety$43–$47/hrMercor · Language & Audio
- Bilingual Hindi Generalist Expert — AI Safety$18–$22/hrMercor · Language & Audio
- Cybersecurity Penetration Testing Expert$60–$120/hrAlignerr · Software Engineering
- AI Security Penetration Tester$60–$120/hrAlignerr · Software Engineering
- AI Red Team Tester$40–$120/hrAlignerr · Business & Operations
- AI Penetration Testing Expert$60–$120/hrAlignerr · Software Engineering
- AI Red Team — Penetration Testing Expert$60–$120/hrAlignerr · Software Engineering
- AI Penetration Tester$60–$120/hrAlignerr · Software Engineering
- AI Red Team Penetration Tester$60–$120/hrAlignerr · Software Engineering
- AI Red Team Analyst$15–$75/hrAlignerr · Software Engineering
- AI Red Team Tester$15–$75/hrAlignerr · Software Engineering
- AI Safety Annotation ReviewerPay not listedTuring · Other
- Lead Penetration Tester$40–$80/hrAlignerr · Software Engineering
- Penetration TesterPay not listedDataAnnotation · Other
- Anthropic Fellows Program, AI Safety & SecurityPay not listedAnthropic · Other
Before you apply
Check each posting before starting an application.
- Read the full posting for duties, screening content, time expectations, and required experience.
- Check the countries the posting names and whether your location matches.
- Check whether listed pay is per hour or per task.
- Keep confidential information out of prompts, files, notes, and reports.
- Never pay to get work.
Quick answers
What is AI red teaming work?
Red teaming tests AI models for unsafe answers within each platform's rules. Many roles review answers, label concerns, and write reports for the platform.
Are there remote AI red teaming jobs?
The board lists 21 open roles for remote AI training work. Country rules vary, and 4 listings name where applicants may live.
Are AI red teaming jobs freelance?
Some roles are freelance or contract roles, while others use different arrangements. Check each posting for the work terms and expected commitment.
How we count
A role is here when its title names red teaming, jailbreaks, adversarial or penetration testing, AI safety, trust and safety, harmful content or content moderation. Patient, food and workplace safety roles are not counted.
Pay is the pay each platform lists. The median counts hourly pay listed as one rate or a range, at its midpoint, from 8 or more roles. We make no estimates and convert nothing. A platform can change a rate or close a role at any time.
Paid research studies and salaried staff jobs are not counted here.
Never pay to get work. An honest platform does not charge you to apply, to train or to be paid.
We counted on .
Back to the open roles. Not sure which one fits you? Match your resume to the open roles. Matching runs in your browser. Your resume stays on your device.
New to this work? Lesson 1 of our free school takes about 5 minutes, and you can get a badge.
More by topic and country: After a layoff · Licensed professionals · For students · Part-time work · Voice and audio · Bilingual work · For nurses · Math · For writers · Entry-level roles in India · All topics
New this week · Entry-level jobs · Jobs by country · Highest pay · Weekly index · All guides · Every role on the board · Every platform · How we count