
QA Speech-to-Text Audio Collection Project - Malay (MY)
TELUS Digital โข Perak, Malaysia
Salary: $5 / hour
-
Role & seniority
-
QA Specialist (Independent Contractor) for Speech-to-Text (STT) script validation
-
No formal seniority level stated; contributor/annotator-type role
-
-
Stack/tools
-
FTS platform for batch script review/QA
-
Other digital QA/language assessment tools (unspecified)
-
Device access (computer/laptop/phone/tablet) + stable internet
-
-
Top 3 responsibilities
-
Validate LLM-generated Malay English (Manglish) scripts for grammatical accuracy, completeness, and naturalness/colloquial accuracy
-
Identify and flag incomplete sentences and critical grammar errors using predefined headers/tags
-
Review batches in FTS platform, flag linguistic inconsistencies/edge cases for escalation
-
-
Must-have skills
-
Native Malay speaker
-
Strong English proficiency, including Malay English/Manglish
-
Deep understanding of Malaysian English colloquialisms/dialect variations
-
High attention to detail; ability to detect grammatical inconsistencies
-
Ability to use FTS platform and perform manual QA
-
-
Nice-to-haves
-
Prior experience in language QA, language assessment, or linguistic work
-
Formal linguistic certification not required
-
-
Location & work type
-
Remote / Work from home
-
Independent Contractor
-
July 2026 โ September 2026, ~20 hours/week
-
Pay: $4.50 USD/hour via Hyperwallet
-
Full Description
Job description
TELUS Digital is seeking quality assurance specialists to participate in a Speech-to-Text (STT) script validation project. Participants will review and validate LLM-generated scripts in Malay English (Manglish), ensuring they meet quality standards for grammatical accuracy, completeness, and naturalness. During script validation reviews through our FTS platform, you will review LLM-generated Malay English (Manglish) scripts with predefined headers/tags, validate scripts for incomplete sentences and critical grammatical errors, conduct manual quality review to ensure naturalness and colloquial accuracy, process and QA sample batches through our FTS platform, and flag linguistic inconsistencies or edge cases for escalation.
The Primary Objective The primary objective of this project is to ensure the integrity and quality of scripts that will be used for audio recording. Your work ensures that only high-quality, natural-sounding scripts are used for audio recording and speech recognition technology development, which will help improve speech recognition models that understand and process natural speech in the Malaysian context.
Expected Project Duration July 2026 โ September 2026
Time to Complete the Task 20 hours per week
Location: Remote (work from home) Qualification path
Payment Rate
- $4.50 USD per hour for script validation tasks
- Payment is calculated based on the hourly tasks assigned and completed. The number of hours and tasks assigned will depend on script availability, your QA performance quality, and project needs.
- This is an Independent Contractor opportunity.
- Payments will be made via Hyperwallet.
Other Requirements
- 18+ years of age
- Native speaker of Malay language
- Proficiency in English (including Malay English/Manglish)
- Strong understanding of Malaysian English colloquialisms and dialect variations
- Keen attention to detail and ability to identify grammatical inconsistencies
- Ability to use FTS platform and other digital QA tools
- Stable internet connection for the duration of the project
- Device (computer, laptop, phone, or tablet) to access the FTS platform
- Prior experience in language QA, language assessment, or linguistic work is preferred (formal linguistic certification not required)
Independent Contractor Relationship This opportunity is offered on an independent contractor basis. Contributors have the flexibility to choose when and how much they work, subject to project availability and quality requirements. This opportunity is not and should not be construed as creating an employment relationship with TELUS Digital.
Contact email datasourcing@telusinternational.ai