Music Production Expert - Russian

What you would do

• Listen to AI-generated songs across diverse genres using professional headphones or studio monitors for critical analysis
• Compare compositions on musicality, creativity, prompt alignment, vocal performance, and production mixing
• Label and categorize songs by genre, structure, instrumentation, and vocal characteristics
• Assess vocal performance and lyrics for authenticity and quality by comparing against professional reference material
• Document feedback in both Russian and English to guide AI model development

Who they want

• Native or near-native Russian proficiency; strong written and spoken English for detailed instructions
• 2+ years' professional background in music production, sound engineering, or mixing work
• Studio-quality headphones or monitors for precise critical listening sessions
• Formal music training in performance, theory, or composition (preferred but not required)
• Familiarity with contemporary Russian music genres, subgenres, and regional artists

Main skills

What the interview asks about

1. 1.AI prompt interpretation and musical intent

   AI models fail when they misinterpret creative direction; your ability to spot misalignment is critical to training better systems.

   For example: “Prompt: melancholic Russian folk ballad, D minor, traditional instruments. AI output: bright uptempo electronic, F major. Categorize this failure. What should the model improve?”

2. 2.Vocal authenticity and emotional expression

   AI-generated vocals often lack human nuance or emotional believability; identifying where they succeed or fail trains models for more convincing singing.

   For example: “An AI generates lyrics in Russian for a love song, but the vocal delivery sounds robotic and over-enunciated, missing the breath and phrasing a human singer would add. How do you rate this and explain what's missing?”

3. 3.Mix quality and production clarity

   Professional-quality AI music must compete with real recordings; poor mixing clarity signals a technical limitation the model needs to overcome.

   For example: “A generated track has good musical ideas but the drums are so loud they mask the vocal melody, and the bass muddies the midrange. Is this a musicality issue or a production issue, and what feedback would improve the model?”

4. 4.Genre authenticity and structural conventions

   Different genres have distinct structural and instrumentation rules; AI must learn these constraints to generate credible output in each style.

   For example: “You're evaluating AI output labeled as Russian folk but it violates traditional modal tonality and rhythm. How do you assess whether this is creative reinterpretation or failure to understand the genre?”

5. 5.Bilingual technical precision

   Feedback must be actionable for engineers and researchers; vague or mistranslated critiques lead to wasted iteration cycles.

   For example: “You need to describe to the development team why a harmonic progression feels unresolved. Write the feedback in Russian and English so both teams understand the exact musical concept you're critiquing.”

A task you may get

Listen to two AI-generated music samples and compare them on musicality, prompt adherence, vocal quality, and production mixing. Provide side-by-side evaluation in both Russian and English, explaining which better matches its original prompt and why.

How to prepare

• Review AI music platforms and examples to familiarize yourself with current capabilities and common failure modes
• Study Russian music terminology and contemporary genre names to ensure precision in labeling and feedback
• Listen to professional reference recordings in your strongest Russian music genres and note what makes them cohesive and authentic
• Practice comparing two musical performances and articulating specific differences in timing, tone, and emotional expression

The facts

Pay

$35–49/hr

Hours

Hourly, 20 hours a week

Where

Remote

Field

Arts & Design

Posted

7/6/2026

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.