AI Safety Expert

Join Wiingy AI's AI Safety Expert Network to work with the frontier AI labs building safer, more reliable, and more trustworthy AI systems. This is an open application for future contract opportunities matched to your background—not a specific job posting. Complete your profile, pass our expert vetting, and you become eligible for AI safety projects as they open. We match experts to work on a rolling basis, so you apply once and stay connected to relevant opportunities as they arise.

Experts in the AI safety network contribute to a range of work depending on the project:

• Evaluating frontier AI models for safety, robustness, reliability, and policy compliance across diverse real-world scenarios.
• Designing adversarial prompts, red-team evaluations, and attack scenarios to identify unsafe model behaviors and failure modes.
• Generating and verifying training data for supervised fine-tuning (SFT), preference feedback (RLHF), and verifiable-reward (RLVR) work.
• Identifying where model reasoning breaks: hallucinations, harmful outputs, policy violations, jailbreak vulnerabilities, or unsafe recommendations.
• Developing safety benchmarks, evaluation rubrics, and gold-standard datasets to measure model performance and alignment.
• Stress-testing AI systems against emerging threats while helping improve model reliability and responsible deployment.

Projects vary in scope and duration, with typical commitments in the range of 5–10 hours per week. Each task is scoped, with its rate shown upfront, and you choose what to take on.

• Bachelor's, Master's, or PhD in Computer Science, Machine Learning, AI, Cybersecurity, Cognitive Science, Philosophy, Public Policy, or a related field, or equivalent professional experience in AI safety or evaluation.
• Strong understanding of AI systems, LLM evaluation, safety risks, alignment techniques, or security testing.
• Excellent analytical communication skills with the ability to document findings, justify safety judgments, and write clear evaluation rationales.
• Ability to work independently and reliably in a remote environment.

Focus areas: AI Safety · LLM Evaluation · Red Teaming · Prompt Injection · Jailbreak Testing · Alignment · Model Robustness · Hallucination Detection · Responsible AI · AI Governance · Safety Benchmarks · Policy Compliance · Risk Assessment · Human Preference Evaluation · Trust & Safety

This is not a specific job posting. By applying, you're joining Wiingy's AI safety talent network.

• Apply once: Complete your profile, upload your resume, and confirm your work location.
• Get verified: We review your credentials and approve qualified experts.
• Stay matched: When projects need AI safety expertise, we match you based on your background and project needs.
• Start working: Get invited to opportunities that fit your schedule and interests, and earn competitive rates working remotely on your own schedule.

• You'll be engaged as an independent contractor; projects can be extended, shortened, or concluded based on need and performance.
• Fully remote, completed on your own schedule.
• Rates are shown upfront per project.
• Payment is made on a per-project or scheduled basis for work delivered.