We are recruiting language speakers worldwide for a special AI Speech Model Initiative: Job 00079910
Source language(s): English Target language(s): All languages
Details of the project: We are currently recruiting language speakers worldwide for a special AI Speech Model Initiative.
Selected participants will assist with multilingual speech data collection and annotation to help improve the speech recognition accuracy of AI models.
I. Work Arrangement
Fully remote; participants may work from home anywhere
Participants must have their own smartphone or computer and a stable internet connection
Recordings must be completed using the app designated by the client
Data annotation must be completed on the designated annotation platform
II. Languages Needed
We are currently recruiting speakers of the following languages:
English — speakers with accents from different regions are welcome
Japanese
Spanish
Portuguese
Indonesian
Thai
Malay
Applicants who are proficient in a regional accent, dialect, or local variety should specify the relevant region when applying.
III. Project Responsibilities
1. Speech Recording
Participants will use the designated AI transcription tool to record authentic, natural conversations based on real life scenarios, including but not limited to:
Casual conversations
Workplace or meeting discussions
Conversations about everyday life
Other authentic and natural conversational scenarios
Recording requirements:
Each participant may record no more than 20 hours in total
Valid recording time may not exceed 2 hours per person per day
Speech must sound natural and reflect an authentic accent
Recordings must be made in a quiet environment
The audio signal-to-noise ratio must be at least 5 dB
Recordings may be completed using a smartphone or computer
2. Data Annotation
Participants will process audio data on the annotation platform. Tasks may include:
Audio segmentation and identification of invalid audio
Speech transcription and ASR text alignment
Speaker identification, segmentation, and labeling
Dialect or language identification
Emotion labeling, such as neutral, positive, and negative
Industry terminology and keyword annotation
Quality assessment, including audio clarity and noise levels
3. Quality Review
After the recording and annotation tasks are completed, the project team will conduct a standardized quality review. Data that does not meet the required standards must be revised or redone according to the feedback provided.
Payment will be calculated only on the basis of valid recording and annotation hours that have passed the final review and acceptance process.
IV. Participant Requirements
Language Proficiency
Proficiency in the language for which you are applying
Native-like, natural, and clear pronunciation
Ability to communicate effectively in everyday situations using the relevant language
Current students or graduates majoring in languages, as well as applicants with experience in translation, foreign-language education, or related fields, will be given preference
Experience
No previous experience is required
The project team will provide instructional videos and written guidelines
Participants may begin official project work after completing the training and passing the qualification test
This job is: already available.
We want to pay for this job: 30 EUR per hour
Who can apply: Freelancers and agencies
Deadline for applying: 12/31/2026
This posting to be kept / deleted: Keep this ad at the site permanently, but disable applying for this job after the deadline
Contact person:[Hidden by TD] Company name:[Hidden by TD] Country: China
IP: 193.218.200.88 (Eskisehir, Turkey) Posted on: Wednesday, 05 Aug 2026, 11:30:28
Number of applications already submitted for this job: 54
If you notice a non-payer or scam here, please let translators know in our
forum.
If you can't find the job announcement you have been notified about or you had seen before, that means the job has been closed already or that posting was unacceptable.
