關於 Kaya Kaya 是 Chemin Sdn Bhd 旗下的 AI 數據標註社群，致力招募優秀人才參與人工智能（AI）項目，並提供長期合作機會。 如果你熟悉香港廣東話、對醫療術語有一定認識，並希望參與真實醫療 AI 項目，我們歡迎你加入！ 工作簡介 我們現正招募具醫療背景或熟悉醫療術語的人士，參與香港醫療語音轉錄項目。 工作主要為聆聽香港醫療場景的廣東話及英文對話，並準確轉錄錄音內容。錄音會自然夾雜廣東話及英文，而英文部分大多為醫療術語，因此具備相關知識將有助於更準確完成工作。…
Audio Transcription Reviewer - Cantonese (Part-time)
Malaysia, Hong Kong
July 22, 2026
Job description
About Kaya
Kaya by Chemin Sdn Bhd is a community of high-performing data annotators helping build the next generation of AI. Join us to work on impactful AI projects, develop your skills, and become part of a growing global community.
About the Role
We're looking for Audio Transcription Reviewers to join our AI language project and help ensure the quality of Cantonese-English transcriptions.
In this role, you'll be part of our Quality Control (QC) team, reviewing completed audio transcriptions to ensure they are accurate, complete, and follow project guidelines before they are delivered to the client.
This project features authentic Hong Kong Cantonese-English conversations, so we're primarily looking for candidates who are fluent in Cantonese and familiar with the Hong Kong accent. While some recordings include healthcare-related conversations and English medical terminology, a healthcare background or familiarity with medical terminology is an advantage, but not required.
If you're detail-oriented, enjoy reviewing language data, and want to contribute to building better AI language technologies, we'd love to hear from you.
Requirements
Project Details
- Project Duration: Approximately 2 months
- Project Period: July – September 2026
- Working Hours: Flexible
- Estimated Workload: Approximately 2 hours per batch
- Onboarding Training: Mandatory (during office hours, GMT+8)
What You'll Be Doing
- Review completed Cantonese-English audio transcriptions for accuracy and completeness
- Verify audio segmentation and speaker labels
- Identify and correct transcription, segmentation, and formatting errors
- Ensure healthcare terminology is transcribed accurately where applicable
- Follow project guidelines to maintain consistent quality standards
- Escalate unclear or complex cases when necessary
- Communicate with the Project Manager for clarification
- Complete assigned batches within the project timeline
Compensation
Quality Control (QC): Earn up to USD 8 for approximately 2 hours of work
Execution Details:
Audio Length: 1–9 seconds
Average Handling time: 1-3 minutes/task
Language Skills
- Fluent in spoken Cantonese and able to read and write Traditional Chinese
- Familiar with the Hong Kong Cantonese accent
- Comfortable reviewing Cantonese-English conversations containing code-switching and occasional English medical terminology
- Fluent in English for project communication and understanding project guidelines
Preferred Qualifications
- Experience reviewing transcriptions, subtitles, or language-related content
- Healthcare background or familiarity with medical terminology is an advantage
- Previous experience in transcription, quality control, data annotation, localization, or AI language projects is a plus
Other Requirements
- Strong attention to detail with the ability to identify transcription and segmentation errors
- Familiar with using a Cantonese Jyutping keyboard
- Able to follow detailed project guidelines consistently
- Access to a laptop or desktop computer, headset, and a stable internet connection
Benefits
Work remotely from anywhere
Priority consideration for future AI projects
Contribute to AI trained on authentic Cantonese-English healthcare conversations
How to Apply
Complete the mandatory pre-assessment as part of the screening process: https://forms.gle/D7kUzwkSgVWW7MvQ7
Available regions
Each region below is open to applicants in that country only. Apply through the link for your region.
Apply now
You will be redirected to the company's website to complete your application.
Welo Data is hiring German-speaking freelancers to help train a leading voice AI platform. The role involves listening to short (5-second) audio clips, accurately transcribing spoken German, and flagging unclear or noisy recordings according to project guidelines.
About Perle 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. Our partners include xAI, Samsung, ELM, and Unisys.
About Rad AI At Rad AI, we’re on a mission to transform healthcare with artificial intelligence. Founded by a radiologist, our AI-driven solutions are revolutionizing radiology—saving time, reducing burnout, and improving patient care.