Amharic Speakers: Audio Transcription and Segmentation Project

About this opportunity

What We're Researching

We're running a paid study to improve the accuracy of Amharic speech recognition models. High-quality audio segmentation and transcription correction feed directly into making these language tools more reliable. This work ensures that automated systems accurately capture the nuances of spoken Amharic.

How It Works

You will log into a dedicated platform to listen to various Amharic audio recordings. Your primary task is to verify and adjust the existing audio segmentation boundaries so they align perfectly with the speech. You will also review the text transcripts provided and correct any errors you hear. This is an ongoing project requiring a minimum commitment of 10 hours per week.

Who This Is For

We are looking for fluent Amharic speakers with strong listening and writing skills. Candidates with a background in transcription, translation, or linguistics are highly encouraged to apply. You must be detail-oriented and able to commit to the ongoing weekly hour requirement.

Requirements:
• Fluent in spoken and written Amharic
• Strong attention to detail when listening to audio recordings
• Comfortable using web-based tools for data annotation and text editing
• Available to dedicate at least 10 hours per week to the project

Responsibilities:
• Listen carefully to provided Amharic audio recordings
• Verify and adjust audio segmentation boundaries to match the spoken words
• Review and correct text transcriptions for accuracy and spelling
• Commit to a minimum of 10 hours per week for ongoing tasks

Check the original posting for the complete requirements and current terms before applying. Applications are handled by TERAC.