Skip to content
SaidGig
Sign up.

Give me your email, I promise I won't do anything weird with it.

Bilingual Marathi Language Evaluator for AI Responses

$15–$20/hr

RemoteContractlanguage
Apply Now

About this role

Role Overview

Assess Marathi AI-generated responses for factual accuracy, reasoning, clarity, tone, and completeness, and produce clear English-language analysis that identifies strengths and specific areas for improvement. Your evaluations will be used to help create the "perfect AI-generated response" in a later phase of the project.

Key Responsibilities
  • Fact-check model outputs using trusted public sources and appropriate external tools.
  • Produce high-quality human evaluation data that documents response strengths, areas for improvement, and factual inaccuracies.
  • Assess reasoning quality, logical consistency, clarity of explanation, tone, and completeness of responses.
  • Verify that model responses align with expected conversational behavior and applicable system guidelines.
  • Write concise, reproducible evaluation artifacts in English that enable consistent downstream improvements.
Qualifications
  • Bachelor''s degree.
  • Native speaker of Marathi, with strong written and spoken English skills.
  • Significant experience using large language models, and a clear understanding of common LLM uses and failure modes.
  • Excellent English writing ability, capable of articulating nuanced feedback clearly.
  • Strong attention to detail, able to identify subtle factual or reasoning issues.
  • Experience in domains that require structured analytical thinking, such as research, policy, analytics, linguistics, or engineering.
Nice-to-Have
  • Prior experience with RLHF, model evaluation, or data annotation.
  • Experience writing or editing high-quality written content.
  • Experience comparing multiple outputs and making fine-grained qualitative judgments.
What Success Looks Like
  • Consistently identifying factual inaccuracies, reasoning errors, and communication gaps in model responses.
  • Delivering clear, consistent, and reproducible evaluation artifacts that others can follow.
  • Your feedback leads to measurable improvements in model response quality and user experience.
Work Terms
  • Remote, global contract position.
  • Engagement is hourly, contractor status.
  • All evaluation notes and deliverables are expected to be written in English.
Compensation
  • 15 - 20 hourly.
Eligibility and Application Process
  • Fluent language requirement: native Marathi and strong English proficiency.
  • To be considered for this role, you must complete the Bilingual Competency interview in Marathi. Candidates who do not take this interview will not be considered.

Related Jobs