Hindi-Speaking AI Session Reviewer (Voice & Video Evaluation)

About the role

We're looking for detail-oriented, native or fluent Hindi speakers to evaluate recorded sessions between users and a voice-and-camera AI assistant. In each session, a user speaks aloud while their camera streams whatever is in front of them, and the AI assistant responds in real time. Your job is to watch these sessions carefully, compare the assistant's spoken responses against what actually appears in the video, and produce clear, structured written assessments using a detailed rubric.

This is judgment-driven work, not high-volume clicking. Your written observations directly inform what gets fixed and prioritized, so accuracy, attentiveness, and clear communication matter more than speed.

What you'll do

• Review recorded sessions where a user interacts verbally with an AI assistant while a live camera feed captures their surroundings
• Carefully compare the assistant's spoken responses to the actual visual content shown in the video
• Identify discrepancies, hallucinations, missed visual cues, or factual errors in the assistant's responses
• Score each session against a structured evaluation rubric
• Write clear, specific, and actionable comments explaining what worked and what failed
• Flag critical failures that should be prioritized for correction
• Maintain consistency and objectivity across many sessions reviewed over time

Requirements

• Fluent or native-level proficiency in Hindi (spoken and written)
• Strong English comprehension for rubric interpretation and written documentation
• Excellent attention to detail and ability to cross-reference audio, video, and text simultaneously
• Strong written communication skills to produce clear, useful feedback
• Comfortable working independently with minimal supervision
• Reliable internet connection and a quiet environment for focused review work
• Prior experience in QA, content moderation, transcription, translation, or annotation work is a plus
• Familiarity with AI assistants or multimodal AI systems is a plus but not required

Compensation

This role pays $12–$18 per hour, depending on experience and demonstrated review quality. Fully remote, flexible schedule, with opportunities for ongoing work based on performance.