Web Research Specialist
Fully Remote | 8-Week Project | Paid in USD
Estimated Earnings: Approximately $30 per approved task
Up to 40 hours/week | Start as soon as the assessment is successfully completed
About the Role
We are looking for experienced Web Research Specialists to contribute to an evaluation benchmark for frontier AI browsing agents.
This is an investigative research role, not a traditional subject matter expert or content-writing position. Your work will focus on designing research problems that are difficult for advanced AI systems to solve, even with full web access and multiple attempts.
You will work from verifiable facts and construct research questions that make those facts exceptionally difficult to locate. Each research problem must be supported by a complete, auditable evidence trail demonstrating how the answer was established.
What You’ll Do
Investigative Web Research
Start from a verifiable fact and work backwards to construct a challenging research question.
Design research problems that are difficult for advanced AI browsing agents to solve.
Investigate unfamiliar subjects from scratch using open-web research techniques.
Locate primary records and authoritative sources across the web.
Research Question Development
Create natural-language research questions with short, stable, and objectively verifiable answers.
Develop clues with specific constraints that require investigation across multiple sources and fact types.
Incorporate facts involving dates, people, places, organizations, works, events, records, and quantities.
Evidence & Source Validation
Build complete and auditable evidence trails supporting research conclusions.
Locate and validate primary records and authoritative documentation.
Navigate government and institutional databases, archives, registries, and PDF documents.
Cite exact pages, tables, sections, and other precise source locations rather than relying on general webpages.
Search Validation
Document the obvious searches performed during the research process.
Record the results returned by those searches.
Demonstrate how the final research problem and answer were validated.
Maintain structured documentation throughout the research process.
AI Evaluation & Benchmarking
Apply experience with LLM evaluation, red-teaming, or benchmark construction where applicable.
Contribute research insights that help evaluate the capabilities and limitations of AI browsing systems.
What You’ll Produce
Your research work will result in:
A natural-language research question with a short, stable, objectively verifiable answer.
Independently checkable clues spanning multiple fact types and specific constraints.
A validation record documenting the obvious searches performed and their results.
A complete evidence trail supporting the research conclusions.
Required Skills
Open-web research
Investigative research
User research
LLM evaluation
Primary-source research
Evidence validation
Precise sourcing
Structured documentation
Research synthesis
Written English
Attention to detail
Minimum Qualifications
Master’s degree OR more than 3 years of relevant experience.
Demonstrated open-web research ability, including locating primary records and navigating government and institutional databases, archives, registries, and PDF documents.
Strong precision with sourcing, including the ability to cite exact pages, tables, and sections rather than general homepages.
Comfort researching unfamiliar subjects from scratch.
Native or near-native written English.
High tolerance for structured documentation, with the understanding that the evidence trail represents the majority of the work.
Experience with LLM evaluation, red-teaming, or benchmark construction.
Experience in one or more of the following areas:
Reference librarianship, archival research, or special collections
Investigative journalism or professional fact-checking
OSINT, due diligence, KYC, or investigative research
Patent, prior-art, or legal-discovery search
Genealogy and records research
Competitive quizzing or puzzle-hunt construction
Nice to Have
Familiarity with JSON and structured data delivery formats.
Project Details
This is an 8-week, fully remote project with paid work in USD. The current task mix provides estimated earning potential of approximately $30 per approved task, with an ideal workload of up to 40 hours per week.
The project can begin as soon as you successfully complete the required assessment.
What Success Looks Like
Success in this role means producing research problems that are genuinely difficult to solve, while maintaining a rigorous and auditable path from the original fact to the final answer.
Strong performance requires investigative curiosity, precise sourcing, disciplined documentation, and the ability to research unfamiliar subjects thoroughly and prove every important conclusion with independently verifiable evidence.
Application Process
Candidates must submit an application and complete the required assessment. Candidates who successfully pass the assessment will be eligible to proceed with onboarding.