AI Benchmark & Red-Teaming Web Research Specialist

You will create research tasks that test where advanced language models struggle. The work combines investigative web research, source checking, benchmark writing, model testing, and review of other tasks.

You will focus on facts that simple searches do not easily reveal. You will use credible sources, document exact URLs, and create precise answer keys that another researcher can defend.

• Research niche topics and identify reliable factual answers.
• Cross-check evidence, resolve contradictions, and record source URLs.
• Write clear benchmark prompts with precise, primary-sourced answer keys.
• Run tasks against language models to test whether they expose the intended weaknesses.
• Review search paths, reasoning outputs, run logs, and source choices.
• Identify whether a failure comes from model limits, search behavior, source selection, or an unclear prompt.
• Apply proven research patterns to new subject areas.
• Review benchmark tasks created by other trainers.

The role is listed as contract and part-time work. Pay details are not provided.

• Time: 20 or more hours per week.
• Location: Open worldwide.
• Language: Fluent in English.
• Experience: At least one year in investigative research, fact-checking, AI benchmark authoring, red-teaming, or technical quality assurance.
• Knowledge: Familiarity with benchmarks such as GAIA, BrowseComp, SimpleQA, SWE-bench, Terminal-Bench, or Humanity's Last Exam.
• Research skills: Intermediate-to-advanced web search and OSINT skills, including work with academic archives, local records, or obscure databases.
• Analysis: Ability to inspect model reasoning, search logs, and source selection.
• Writing: Clear technical writing for unambiguous prompts and defensible answer keys.
• Education: A relevant bachelor's degree or higher in an engineering field.
• Helpful experience includes prompt-injection or reasoning-trap tasks, published investigative or academic research, difficult-to-source records, and benchmark review or auditing.

Apply on OpenTrain with your resume, then complete the application on the hiring site.

AI training is the human work behind systems that learn from examples, including writing and rating model responses, evaluating outputs, and testing search behavior. Research specialists help make these examples accurate and challenging so evaluation results show how models perform in realistic situations.