Safety Specialist (LLM) | Remote

Crossing Hurdles; United Kingdom; On-site; GBP 60,000 - 80,000; Full time

Job summary

A technology evaluation company in the United Kingdom seeks a specialist to evaluate frontier AI systems. Responsibilities include annotating model failures, identifying vulnerabilities, and producing actionable insights. Qualified candidates will have a strong adversarial mindset and clear communication skills for diverse audiences. The role requires independence in a fast-paced remote environment, focusing on systemic risks and safety frameworks.

Qualifications

• Ability to push AI systems to their limits.
• Experience in applying structured evaluation frameworks.
• Comfortable communicating with both technical and non-technical audiences.

Responsibilities

• Annotate model failures and classify vulnerabilities.
• Identify systemic risks and document attack cases.
• Produce clear reports and actionable insights.

Skills

Adversarial mindset

Clear written communication

Experience with evaluation frameworks

Ability to operate across safety domains

Independent work in remote setting

Job description

• Annotate model failures and classify vulnerabilities in frontier AI systems.
• Identify systemic risks and document reproducible attack cases.
• Apply structured taxonomies, benchmarks, and safety frameworks.
• Produce clear reports, datasets, and actionable insights for customers.
• Support diverse projects including jailbreak testing and socio technical abuse evaluations.

Requirements

• Strong adversarial mindset with the ability to push systems to breaking points.
• Experience applying structured evaluation frameworks and benchmarks.
• Clear written communication skills for technical and non technical audiences.
• Ability to operate across multiple safety domains and evolving project scopes.
• Comfort working independently in a fast paced remote environment.