Xin Yi Ooi

Key Skills

Software

Appen

Labelbox

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Document

Text

Top Task Types

Audio Recording

Classification

Evaluation Rating

Translation Localization

Freelancer Overview

As an AI Trainer, I specialized in the critical data labeling and annotation of audio data for a low-resource language: Malaysian Cantonese. My core responsibility involved labeling audio data, listening to short, often challenging social media audio clips and producing accurate, nuanced transcriptions in Malaysian-Yue Chinese. This work required not just linguistic skill but also a deep cultural understanding to capture local slang, expressions, and phonetic variations unique to the dialect, directly contributing to the AI's ability to parse and understand a specialized form of human communication. What distinguishes my contribution is the focus on a specific linguistic niche. This project demanded more than general transcription skills; it required expertise in a particular dialect's colloquialisms and a reviewer's meticulous eye for quality assurance. By ensuring the highest accuracy in my transcriptions and evaluating the work of others, I played a direct role in building a high-quality, culturally-aware training dataset. This experience has honed exceptional skills in linguistic precision, cultural fluency, and a rigorous quality evaluation process essential for developing robust and inclusive AI models.

Labeling Experience

Native Mandarin Audio Transcriber – Simplified Chinese

LabelboxAudioClassificationText Generation

- Accurately transcribed short Mandarin audio clips into Simplified Chinese, meticulously capturing all speech exactly as spoken - Ensured strict adherence to detailed style guides for formatting, spelling, numbers, and multispeaker handling, and applied project-specific annotation conventions - Consistently delivered high-quality transcripts and conducted peer reviews to perform quality control, identifying and correcting errors in spacing, formatting, and tags - Evaluated and selected the highest-quality transcript from peer submissions, prioritizing accuracy while preserving natural wording variations in line with client objectives.

- Accurately transcribed short Mandarin audio clips into Simplified Chinese, meticulously capturing all speech exactly as spoken - Ensured strict adherence to detailed style guides for formatting, spelling, numbers, and multispeaker handling, and applied project-specific annotation conventions - Consistently delivered high-quality transcripts and conducted peer reviews to perform quality control, identifying and correcting errors in spacing, formatting, and tags - Evaluated and selected the highest-quality transcript from peer submissions, prioritizing accuracy while preserving natural wording variations in line with client objectives.

2025 - 2025

Cantonese Transcription in Malaysia

AppenAudioEntity Ner ClassificationClassification

- Annotated, transcribed, and validated large volumes of conversational audio data in Mandarin and Cantonese - Applied linguistic knowledge and cultural context to capture natural speech, including slang and dialect - Performed quality control reviews on peer submissions, maintaining project accuracy above 95% - Adapted to changing project guidelines while consistently meeting deadlines - Collaborated within a remote team environment, contributing to AI model training and improvement

- Annotated, transcribed, and validated large volumes of conversational audio data in Mandarin and Cantonese - Applied linguistic knowledge and cultural context to capture natural speech, including slang and dialect - Performed quality control reviews on peer submissions, maintaining project accuracy above 95% - Adapted to changing project guidelines while consistently meeting deadlines - Collaborated within a remote team environment, contributing to AI model training and improvement

2025 - 2025

Education

Work History