Senior GenAI Safety Researcher
Posted 5ds ago
Employment Information
Report this job
Job expired or something wrong with this job?
Job Description
Senior GenAI Safety Researcher evaluating vulnerabilities across generative AI systems. Designing red-team methodologies for Alice, a trust, safety, and security company.
Responsibilities:
- Architect rigorous, scalable testing methodologies and red-teaming frameworks for foundational models, multimodal systems, and AI agents
- Develop prompt strategies across risk domains to expose complex model vulnerabilities
- Research emerging jailbreak tactics, prompt injection techniques, and circumvention strategies
- Provide technical and policy insight to support the Program Lead in scoping projects, assessing risks, and driving strategy
- Document, synthesize, and expand Alice’s internal AI Safety knowledge base
- Standardize best practices, taxonomies, and research findings across the team
- Mentor junior analysts and elevate the team’s analytical standards
- Own engagement lifecycles from planning and methodology design through execution, QA, and final delivery
- Oversee complex, multi-language datasets across multiple areas of abuse
- Partner with engineering, product, policy, and client-facing teams to communicate findings and inform mitigation strategies
Requirements:
- 5+ years of experience in AI Safety, Responsible AI, Trust & Safety, or aligned research domains
- Proven expertise in research design and building qualitative or quantitative evaluation methodologies for GenAI
- Strong domain expertise in content risks, including toxicity, copyright, misinformation, and safety policy violations
- Track record of project ownership, leading deliverables end-to-end with high attention to detail in fast-paced, variable environments
- Deep familiarity with modern Generative AI architectures, prompt engineering, red-teaming, and AI agents
- Strong communication skills to act as a core subject-matter contact for program leads, internal teams, and clients
- Published research in academia, industry whitepapers, or a research institute is nice-to-have
- Hands-on experience evaluating multimodal systems is nice-to-have
- Experience mentoring, leading, or QAing junior analysts and researchers is nice-to-have







