$30–45/hr · micro1
A music annotation expert who documents audio characteristics from real-world recordings and evaluates AI model accuracy.
What you would do
- Listen to field recordings and complete detailed annotations identifying instruments, tempo, key, and ambient sounds
- Use a structured web platform to mark event timestamps and categorize audio elements
- Compare paired recordings of the same music from different locations and describe what changed
- Review AI model descriptions of audio clips and score them for accuracy, partial correctness, or errors
- Provide brief written explanations supporting your assessment of each annotation
Who they want
- 3+ years of hands-on experience in music, audio production, sound design, or music education
- Strong analytical listening that isolates individual instruments from busy real-world mixes
- Ability to estimate tempo and identify musical key, including with relative pitch methods
- Comfort with web-based annotation tools, dropdowns, timestamps, and structured note-taking
- Clear written English for concise, evidence-based explanations (20-30 hours per week availability)
Main skills
What the interview asks about
1.Instrument isolation in noisy fields
Studio recordings are clean; field work means traffic, crowd noise, and weather. You need to extract information from chaos.
For example: “You hear a violin in a recording with heavy traffic noise in the background. How confident are you identifying it, and what details convince you it's there?”
2.Tempo determination under variability
Live or outdoor recordings often have tempo fluctuation. Your estimate needs to be realistic within a few BPM and justify your reasoning.
For example: “This live recording drifts slightly in tempo, speeding up during an energetic section. What tempo would you report and how would you note the variation?”
3.Comparative analysis across conditions
The same song in two locations will sound different. You must judge what's essential to the music versus what's context-dependent.
For example: “You have street recordings of a busker playing the same song at two different times. What musical elements stayed identical and what shifted based on environment?”
4.AI grading with evidence
Simply marking an AI response as wrong is useless; you need to explain exactly what was missed, exaggerated, or invented.
For example: “An AI says there's a bassline starting at 0:15, but you hear it begin at 0:18. How do you document this discrepancy concisely?”
5.Form and scale accuracy
Annotation platforms have specific fields and formats. Sloppy entry ruins downstream training, so precision in data entry is as important as listening skill.
For example: “You need to log four different background sounds at specific timestamps using dropdown categories. What process do you use to avoid mixing up entries?”
A task you may get
Annotate a 3-minute field recording identifying all instruments, estimating tempo, noting background sounds with timestamps, and scoring a sample AI description of the clip.
How to prepare
- Listen to 10-15 field recordings and try identifying instruments, tempo, and key before checking reference answers
- Watch recordings of live performances in various outdoor settings to train ear for real-world complexity
- Practice using web annotation tools with sample audio to develop speed and minimize entry errors
- Review examples of good and poor AI audio descriptions to understand what grading rubric you'll apply
The facts
- Pay
- $30–45/hr
- Open to
- Bangladesh, Hong Kong, India, Indonesia, Japan, Kazakhstan, Kyrgyzstan, Malaysia, Pakistan, Philippines, Singapore, Sri Lanka, Taiwan, Thailand, Uzbekistan, Vietnam, Austria, Belarus, Belgium, Denmark, France, Germany, Greece, Italy, Netherlands, Portugal, Russia, Spain, Switzerland, United Kingdom, Argentina, Brazil, Chile, Colombia, Mexico, Peru, Algeria, Bahrain, Egypt, Iraq, Jordan, Kuwait, Lebanon, Libya, Morocco, Oman, Palestine, Qatar, Saudi Arabia, Tunisia, United Arab Emirates, United States, Canada, Nigeria, Kenya, South Africa, Ghana, Ethiopia
- Field
- Language Audio
- Role type
- Specialist
- Posted
- 9/2/2026
- Places left
- 10
We wrote this page from the public micro1 listing. It may be out of date, so read the full posting before you apply.