Perle is an AI infrastructure company building expert-driven training data, evaluation systems, and applied AI products for the world's leading labs and enterprises. Headquartered in San Francisco with experts across more than forty markets, we specialize in the work that requires real human judgment: domain expertise, linguistic nuance, and cultural fidelity that generic data vendors cannot deliver.
Perle is hiring a freelance transcriptionist fluent in Sinhalese for a remote transcription role. This is a paid contract position open to qualified contributors worldwide.
- Role Overview
- What You Will Do
- Listen to 2-speaker audio recordings in Sinhalese
- Type out what each speaker says, accurately and in full
- Label each speaker turn as Speaker 1 or Speaker 2, matching the audio channel
- Add timestamps at the start of each speaker turn (format: HH:MM:SS)
- Ensure transcripts are sequential with no unexplained gaps
- Requirements
- Native or near-native fluency in Sinhalese (spoken and written)
- Strong written Sinhalese â accurate spelling, punctuation, and grammar
- Ability to work independently and meet daily output targets
- Access to a computer with a stable internet connection
- Availability for at least 4 hours per day
- Prior transcription experience is a plus but not required â all contributors must pass a qualification test
- Quality Standard
- Word Error Rate (WER) of 5% or less on all delivered transcripts
- Zero speaker label errors â Speaker 1 must always correspond to Channel 1
- Transcripts reviewed on a rolling basis; feedback provided promptly
- Files that do not meet the quality threshold will be returned for correction
- Project Details
Information
Language required
Sinhalese
Project type
Audio transcription (manual)
Commitment
~at least 4 hours/day (remote, flexible schedule)
Start date
As soon as qualified â rolling onboarding
- How to Apply
Perle is an equal opportunity employer. We welcome applicants from all backgrounds.
Please mention the word **FANTASTIC** and tag RMTg1LjYxLjE1NS40Mg== when applying to show you read the job post completely (#RMTg1LjYxLjE1NS40Mg==). This is a beta feature to avoid spam applicants. Companies can search these words to find applicants that read this and see they're human.