Music & Lyrics Expert - Punjabi

What you would do

• Compare AI-written lyrics against published Punjabi songs to spot imitation and derivative content
• Rate lyrics on overall quality, creative originality, and adherence to creative prompts
• Evaluate natural delivery: word choice, slang, regional expressions, and cultural nuance
• Assess whether lyrics match different Punjabi genres, sub-genres, and artists
• Provide detailed feedback on model strengths and failure patterns

Who they want

• Native or near-native Punjabi language proficiency with ability to read and follow English instructions
• Background as songwriter, lyricist, performer, composer, or music journalist familiar with Punjabi music
• Broad familiarity with contemporary Punjabi genres and sub-genres
• Strong ear for authenticity, cultural expression, and linguistic patterns
• Quality headphones or studio monitoring equipment suitable for detailed audio listening

Main skills

What the interview asks about

1. 1.Lyric authenticity and natural delivery

   The model must learn what sounds like a real Punjabi song; distinguishing AI stiffness from organic flow is your core skill.

   For example: “An AI system generates the line 'Main tere dil di khabar vich rau haan' for a bhangra track. Compare this to how a Punjabi lyricist would phrase a similar emotion. What's missing or off about the AI version?”

2. 2.Genre-specific convention recognition

   Punjabi music spans folk ballads to rap; the model must learn each sub-genre's unique style and vocabulary.

   For example: “You're rating AI lyrics for a traditional Sufi-style Punjabi poem versus a modern Punjabi rap verse. How would your quality criteria differ between these two?”

3. 3.Originality judgment and plagiarism spotting

   AI models sometimes memorize training data; you must catch when output resembles existing published work too closely.

   For example: “An AI generates a verse that sounds very familiar. You recall it's structurally similar to a well-known Punjabi film song from 2015, though the exact words differ. Is this a failure, and why?”

4. 4.Cultural and linguistic nuance

   Punjabi slang, idioms, and cultural references won't make sense to systems trained on generic song data.

   For example: “The AI generates a lyric using village-specific slang that's authentic but hyper-local and would confuse most modern listeners. Does this affect how you rate creativity versus authenticity?”

5. 5.Prompt adherence assessment

   The model must learn to follow creative direction; rating compliance shows whether it understood the constraints.

   For example: “The prompt asks for an upbeat, wedding-celebration style song with imagery of flowers and family gatherings. The AI output is melancholic and focuses on lost love. How do you score this?”

A task you may get

Review two AI-generated Punjabi song snippets and one published song, rating each on originality, cultural authenticity, and prompt adherence with brief explanations.

How to prepare

• Listen to Punjabi songs across multiple genres (folk, devotional, film, rap, pop) and note stylistic and linguistic patterns
• Read about Punjabi music scene developments and emerging artists to refresh your familiarity with contemporary trends
• Review a few existing lyric annotations or quality rubrics to understand what makes strong versus weak output ratings

The facts

Pay

$15/hr

Hours

Hourly, 20 hours a week

Where

Remote

Field

Language and Audio

Posted

8/12/2026

Places left

1

We wrote this page from the public Mercor listing. It may be out of date, so read the full posting before you apply.