Web Research Specialist

About the Role

We are looking for an AI Research Evaluator to design difficult, verifiable research problems for frontier AI browsing agents.

This is an investigative research rolenot a conventional subject-matter expert or content-writing position. You will begin with a verifiable fact, work backwards to create a challenging research question, and build a complete evidence trail proving the answer.

• Design complex, natural-language research questions with objectively verifiable answers.
• Develop interconnected clues involving people, dates, places, organisations, works, events, records, and quantities.
• Research government databases, archives, registries, institutional websites, PDFs, and primary records.
• Validate information using reliable and independently checkable sources.
• Document exact pages, tables, sections, URLs, and supporting evidence.
• Record obvious searches attempted and explain why they did or did not reveal the answer.
• Prepare structured research outputs for AI benchmark and evaluation datasets.
• Identify weaknesses, search limitations, and failure points in AI-generated research.

• 3+ years of professional experience in open-web, investigative, archival, factual, or records-based research.
• Demonstrated ability to locate and verify primary sources.
• Strong written English communication skills.
• Excellent attention to detail and source precision.
• Ability to research unfamiliar topics independently.
• Experience with LLM evaluation, AI red-teaming, benchmark creation, or fact-checking.
• Comfort creating detailed, structured evidence documentation.
• Familiarity with JSON or structured data formats is desirable.

• Remote contractor assignment
• 40 hours per week
• Minimum 4 hours of PST overlap required
• Contract duration: 8 weeks
• No medical benefits or paid leave under the contractor arrangement