AI Benchmark & Red-Teaming Specialist
Turing
Verified listingOverview
Turing is a research accelerator for frontier AI labs that helps enterprises deploy advanced AI systems. This role involves building evidence-based benchmark tasks designed to expose weaknesses in state-of-the-art AI models through rigorous research, testing, and failure diagnosis.
What You’ll Do
- Discover obscure, hard-to-surface facts through deep research on niche topics
- Verify facts rigorously against primary or highly credible sources and cross-check for contradictions
- Draft clear, answerable benchmark tasks designed to expose structural weaknesses in model research strategies
- Execute tasks against SOTA models to confirm they reliably trigger intended failures
- Trace model run logs, search trajectories, and intermediate outputs to diagnose exact failure points
- Apply validated model-breaking research patterns across new domains and review peer-authored tasks
Requirements
- 1+ years in investigative research, fact-checking, AI benchmark task authoring, red-teaming, or technical QA review
- Proven experience with AI benchmarks (GAIA, BrowseComp, SimpleQA, SWE-bench, Terminal-Bench, or Humanity's Last Exam)
- Intermediate-to-advanced OSINT and web research skills
- Strong analytical capability to inspect model reasoning chains and diagnose failure root causes
- Ability to author clear, precise benchmark prompts with primary-sourced answer keys
- Bachelor or higher degree from relevant engineering domain
Who Should Apply
This role suits someone who is analytically obsessive, enjoys deep investigative work, and gets satisfaction from breaking systems through rigorous evidence—someone who can spend hours tracing model reasoning chains to understand failure mechanisms. It will frustrate people who prefer straightforward, well-defined tasks, who dislike ambiguity in source verification, or who are uncomfortable with the precision required to write unambiguous ground-truth answers for edge cases.
Salary Insight
No salary is stated in the posting. For contract work of this technical nature requiring advanced research skills, you should ask about compensation structure early—whether it's hourly, per-task, or a flat engagement rate. Research similar AI evaluation and red-teaming contract roles before the conversation to anchor your expectations appropriately.
About Turing
Remote work platform connecting skilled professionals with AI labs and companies for software engineering, AI training, data, and specialist contract work.
Remote policy: Remote worldwide, but location eligibility varies by individual project. Many AI roles require minimum hours per week and U.S. Pacific Time overlap. Contractors and freelancers only, not employees.
How hiring works at Turing →Required Skills
Compensation
Not disclosed
- Location
- Worldwide (4 hours overlap with PST required)
- Engagement
- Contract
- Posted
- Yesterday
Opens Turing’s listing on MyRemoteJobs — we don’t collect applications ourselves.
About Turing
Remote work platform connecting skilled professionals with AI labs and companies for software engineering, AI training, data, and specialist contract work.
Remote policy: Remote worldwide, but location eligibility varies by individual project. Many AI roles require minimum hours per week and U.S. Pacific Time overlap. Contractors and freelancers only, not employees.
How hiring works at Turing →Will your CV get past the filter?
Most applications are rejected by software before a person ever reads them. Scan your CV using our ATS Checker tool, see exactly what's blocking it, then fix it in the CV Builder and apply knowing it'll get through.
Scan my CVGet roles like this by email
Tell us what you do and we'll email you when a similar remote job is posted. No account needed.
See our jobs more often in your search
Add MyRemoteJobs as a Preferred Source on Google. Free, takes seconds, and you stay on this page.
Application Tip
Lead with a specific benchmark task you've authored or reviewed, or a red-teaming project where you identified a real model failure mechanism—not just a list of benchmarks you know. Emphasize concrete examples of your OSINT work: a fact you uncovered that initially appeared impossible to find, or a source verification process you designed. This role requires evidence that you can diagnose *why* a model fails, not just that it does; your strongest applications will include a short case study of a failure path you traced end-to-end.
Sourced from MyRemoteJobs · verified against the original posting
Similar open positions
Polish Bilingual Expert
micro1
Verified
Search Generalist
Mercor
VerifiedAnonymization Writer
Mercor
Verified