AI Safety Experts — English & Swedish

Apply Now on Mercor →

The Rise of AI Safety: Why Human Red Teaming Matters More Than Ever

Artificial intelligence is evolving at a breakneck speed, reshaping industries, creativity, automation, and global communication. However, as large language models (LLMs) and conversational AI agents become more deeply integrated into enterprise operations and daily human workflows, the demand for robust safety mechanisms has never been more critical. Automated scripts, code scanners, and static checks can only go so far when it comes to predicting how creative human prompts might destabilize, bias, or manipulate a cutting-edge machine learning model.

This is precisely why human-driven AI red teaming has emerged as one of the most exciting, high-paying, and intellectually demanding frontiers in modern technology. Companies building the next generation of foundational models are aggressively seeking bilingual and multilingual experts who can think like adversaries, probe systemic vulnerabilities, and stress-test conversational models before they ever reach public deployment.

If you possess native-level fluency in both English and Swedish, combined with an analytical mindset, a knack for adversarial problem-solving, and a passion for technology, Mercor’s newly opened remote contract role—AI Safety Experts — English & Swedish—presents an extraordinary career opportunity to get paid handsomely while helping shape the future of safe, trustworthy artificial intelligence.

Overview of the Position and Compensation Structure

Flexibility meets high-tier compensation in this unique remote contracting role. Designed for independent professionals who want to leverage their linguistic and analytical mastery on their own schedule, this engagement offers top-tier hourly rates:

• Pay Rate: $48 – $62 per hour, paid weekly via Stripe or Wise.
• Work Format: 100% Remote, flexible hours completed entirely on your own schedule.
• Contract Type: Hourly independent contractor engagement.
• Required Languages: Native or near-native fluency in English and Swedish.

Whether you are looking to supplement your income, transition into the booming tech and AI governance sector, or apply your unique cross-lingual communication skills to cutting-edge tech projects, this role provides the ideal blend of autonomy and competitive financial rewards.

Detailed Job Description: What Will You Actually Be Doing?

At Mercor, the core philosophy behind AI safety is simple: the safest AI model is the one that has already been rigorously attacked and audited by human experts. As an AI Safety Expert focusing on English and Swedish language data, your primary mission will be to act as an ethical hacker and linguistic auditor for conversational agents.

The role involves reviewing AI-generated outputs that touch upon complex or sensitive topics, including potential biases, cultural nuances, misinformation risks, or harmful conversational trajectories. All tasks are text-based, and participation in sensitive projects is fully transparent, voluntary, and supported by robust psychological wellness resources and clear guidelines.

Key Responsibilities:

• Adversarial Red Teaming: Systematically push conversational AI models and multi-turn agents to their absolute breaking points using jailbreaks, prompt injections, complex scenario manipulation, and edge-case exploits.
• Human Data Generation: Produce high-grade human annotations detailing model failures, classify systemic vulnerabilities, and document anomalous behavior patterns across both English and Swedish datasets.
• Strict Taxonomy Adherence: Apply structured evaluation frameworks, established benchmarks, and rigorous testing playbooks to maintain absolute consistency across all evaluation cycles.
• Comprehensive Documentation: Craft clean, reproducible incident reports, datasets, and attack playbooks that AI developers and enterprise clients can immediately act upon to patch security holes.

Who Qualifies? Profile of an Ideal Candidate

Mercor looks for sharp, curious, and structured thinkers who possess a rare combination of linguistic precision and adversarial creativity. You do not necessarily need a traditional background in software engineering, though technical familiarity is always welcome. The ideal profile includes:

• Prior Red Teaming or Security Background: Experience in AI adversarial testing, cybersecurity penetration testing, or socio-technical probing.
• An Inquisitive & Adversarial Mindset: A natural instinct to question assumptions, find loopholes, and push complex systems past their intended operating parameters.
• Methodical and Structured: Ability to rely on structured frameworks, reproducible testing criteria, and analytical benchmarks rather than random trial-and-error.
• Strong Communication Skills: Capability to clearly articulate abstract systemic risks, behavioral biases, and security vulnerabilities to both technical engineers and business stakeholders.
• High Adaptability: Comfort with moving fluidly across changing project scopes, evolving model architectures, and diverse client requirements.

Valued Specializations & Nice-to-Haves:

If you have niche experience in any of the following areas, your application will stand out significantly during the review process:

• Adversarial Machine Learning: Familiarity with jailbreak datasets, prompt injection techniques, reinforcement learning from human feedback (RLHF/DPO) attacks, and model extraction.
• Cybersecurity Foundations: Practical background in vulnerability research, penetration testing, or software reverse engineering.
• Socio-Technical Risk Analysis: Experience analyzing disinformation, online harassment, conversational safety guidelines, or socio-cultural bias in digital media.
• Creative Probing: Background in psychology, creative writing, theater/acting, or roleplay design to simulate unconventional human deception patterns.

Why Join the Mercor Ecosystem?

Mercor is revolutionizing how top-tier global talent connects with pioneering artificial intelligence labs and enterprises. By bypassing archaic, multi-round interview loops, Mercor utilizes modern, streamlined evaluation workflows that respect your time and highlight your true professional merit.

As a contractor on this platform, you gain direct exposure to the frontier of machine learning development. You collaborate with world-class researchers and engineers while earning top-market compensation ($48–$62/hr) paid promptly and reliably on a weekly basis.

Step-by-Step Guide to Completing Your Application

To maximize your chances of acceptance and secure your spot as an early applicant for this high-demand role, follow these essential application steps carefully:

1. Prepare Your Profile: Ensure your resume or professional profile clearly highlights your linguistic background (specifically English and Swedish proficiency), any analytical or security-related testing experience, and your problem-solving capabilities.
2. Complete the Initial Screening: Click any of the application links on this page to visit the official Mercor job portal.
3. Finish the Entire Application Flow: Do not abandon the process midway. Complete all required verification steps, questionnaire prompts, and interview workflows. Providing thorough and accurate responses ensures your profile lands at the top of the evaluation queue.
4. Await Swift Onboarding: Mercor’s streamlined evaluation process means successful candidates move through review much faster than traditional corporate pipelines. Keep an eye on your email and dashboard for next steps!

Complete Your Full Application Now →