Spanish (Argentina) Linguistic Expert at Mercor

Spanish (Argentina) Linguistic Expert | $40/hr | Remote | Part-time

Mercor is hiring Spanish (Argentina) linguistic experts for an AI research project run in partnership with a leading research lab. The work focuses on evaluating how well a large language model (LLM) performs in Argentine (Rioplatense) Spanish and on curating high-quality Spanish linguistic data to train and improve the model's understanding and generation capabilities.

What the work involves

• Evaluating the quality and accuracy of LLM-generated Spanish text across use cases such as translation, summarization, dialogue, and grammar correction.
• Editing, refining, or rewriting Spanish content to meet linguistic and cultural standards for Argentina.
• Annotating Spanish text for grammatical, syntactic, and semantic features.
• Identifying model errors and language gaps, and collaborating with AI researchers to improve LLM performance.
• Providing detailed linguistic feedback and analysis to guide ongoing model training efforts.

Who this role is for

• Native or near-native fluency in Spanish (Argentina), with excellent command of grammar, syntax, and stylistics.
• Professional experience in translation, editing, linguistic annotation, or NLP data work in Spanish.
• Strong understanding of Argentine (Rioplatense) Spanish nuances, regional variations, register, and context-appropriate usage, including how it differs from other Latin American and Peninsular Spanish varieties.
• Professional fluency in English.
• Interest in AI and LLM-driven language technologies (preferred, not required).
• Excellent attention to detail and the ability to spot subtle errors or unnatural phrasing.
• Clear communication skills and experience working in distributed, remote teams.

Pay and schedule

The role pays $40/hr and is a part-time, remote position. You'll contribute to AI research in partnership with a leading research lab, with hands-on exposure to state-of-the-art LLM development where your linguistic expertise and judgment directly shape model quality.