AI Safety Red Team Analyst for Model Evaluation

OpenTrain AI, Inc.; United Kingdom; Remote; GBP 12,000 - 24,000; Part time

Benefits offered by this job

Remote work

Self-set schedule

Flexible hours

Job summary

OpenTrain AI, Inc. is seeking a curious, entry-level Safety Validator to test how large language models respond to single-turn image-edit requests.

You will classify prompts and outputs according to project guidelines, review ambiguous or benign requests, and write precise rationales to support each decision. This is a part-time, contract engagement starting immediately, with 20+ hours weekly for 1–2 weeks.

Qualifications

• Strong analytical judgment.
• Ability to articulate clear rationales for classifications.
• Experience or familiarity with AI safety concepts.

Responsibilities

• Classify image-edit prompts and outputs per guidelines.
• Review ambiguous, borderline, and benign requests consistently.
• Write precise rationales supporting each classification decision.
• Document bypassed safeguards and subtle policy violations.
• Create adversarial prompts and compare model outputs.
• Identify unclear or conflicting guidelines and suggest clarifications.

Skills

Policy analysis

Prompt classification

Rationale writing

AI safety concepts

Adversarial prompts

Education

Policy, law, ethics, linguistics, journalism, or computer science

Job description

OpenTrain AI, Inc. is seeking a curious, entry-level Safety Validator to test how large language models respond to single-turn image-edit requests.

You will classify prompts and outputs according to project guidelines, review ambiguous or benign requests, and write precise rationales to support each decision. This is a part-time, contract engagement starting immediately, with 20+ hours weekly for 1–2 weeks.