Linguistics Expert for AI Model Evaluation (English)

The work

Create language-analysis tests and assess how accurately AI models reason about linguistic data.

• Write scenarios testing syntactic ambiguity, semantic entailment, and phonological patterns.
• Prepare reference analyses that explain the evidence, generalizations, exceptions, and reasoning behind each judgment.
• Rate model responses for analytical accuracy and appropriate handling of linguistic data.
• Flag responses that misparse a sentence or misidentify the linguistic phenomenon.

What it pays and takes

• Pay: $50 to $100 per hour, paid weekly via Stripe.
• Schedule: Ongoing, asynchronous work requiring at least 20 hours per week, with flexibility to scale your workload.
• Location: Remote; applicants must be based in an eligible country listed for the role.
• Work type: Part-time contractor work.
• Requirements: Strong written English and depth in a linguistics area such as syntax, semantics, phonology, morphology, sociolinguistics, or psycholinguistics.
• You should be able to explain linguistic judgments clearly and identify what evidence could disprove an analysis.
• Helpful background: PhD-level training, peer-reviewed publications, computational linguistics or corpus methods, or fieldwork and documentation of understudied languages.

How it works

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

About AI training work

AI training uses examples and judgments prepared by people to help build and evaluate AI systems. Linguistics expertise helps make those evaluations precise and explains where a model’s reasoning succeeds or fails.