Generalist - English & Hebrew [Remote]

Generalist - English & Hebrew is a remote Hebrew specialist track for evaluating hebrew evaluation outputs against native-speaker standards. Reviewers spot fluency, register, and cultural-context errors that automated checks miss, and write structured rationale that the modeling team can act on in the next training cycle.

Why this role matters

Multilingual model quality depends on reviewers who can read Hebrew the way a native speaker does. AuraOne uses these reviewers to set the bar for tone, formality, and cultural register before a model ever ships to Hebrew users.

Responsibilities

• Evaluate Hebrew model responses for fluency, accuracy, register, and cultural fit for Generalist - English & Hebrew assignments.
• Compare paired hebrew evaluation outputs and pick the stronger response with a written rationale.
• Flag hallucinations, code-switching errors, and formality mismatches with task-specific severity tags.
• Capture native-speaker edits that demonstrate the correct phrasing alongside each issue.
• Calibrate against the AuraOne Hebrew rubric in weekly reviewer-quality cycles.
• Surface ambiguous prompts back to the program team so the rubric can be sharpened.
• Maintain reviewer notes that document tricky idioms, regionalisms, and emerging usage.

Qualifications

• Native or near-native Hebrew fluency with strong written-English ability for Generalist - English & Hebrew work.
• Demonstrable experience editing, translating, or evaluating Hebrew content.
• Comfort applying multi-page rubrics consistently across long evaluation batches.
• Clear written reasoning that names the issue and the standard being applied.
• Reliable async availability for at least 10 hours per week.
• Prior model-evaluation, annotation, or human-rater experience is a plus.

Example tasks

• Score a pair of Hebrew responses to a customer-support prompt and pick the better one with a 1-2 sentence rationale.
• Flag formality mismatches in hebrew evaluation outputs (T/V usage, honorifics, regional register).
• Rewrite a model's Hebrew response so it reads naturally and tag the original error category.
• Audit a 50-row evaluation batch for rubric consistency and surface drift back to the program lead.

Nice to have

• Background in Hebrew linguistics, journalism, or translation studies.
• Experience with TMS tools, MQM error taxonomies, or other structured QA frameworks.
• Familiarity with LLM evaluation rubrics and inter-rater agreement workflows.

Skills

• Multilingual evaluation
• Native-speaker review
• Rubric application
• LLM output evaluation
• Hebrew evaluation

Work model

Remote — US-eligible. Remote · Independent specialist contractor. Employment type: CONTRACTOR. Applicants must be authorized to work from US.

Compensation

$41 / hr

Application process

Apply through AuraOne's specialist intake for role-specific routing and review. Final project scope, schedule, and contractor terms are confirmed before placement.