Junyi J.

Key Skills

Software

Other

Top Subject Matter

Multilingual speech/voice audio for AI training

Top Data Types

Audio

Video

Image

Top Task Types

Audio Recording

Classification

Question Answering

RLHF

Transcription

Evaluation/Rating

Data Collection

Text Generation

Freelancer Overview

I am a third-year Advertising undergraduate and independent audio producer with hands-on experience in multilingual voice recording and AI audio quality evaluation. My work focuses on assessing AI tutor-style audio capabilities, including pronunciation accuracy, tonal consistency, and natural conversational flow across Cantonese, Mandarin and english. My music production background gives me a strong eye for audio detail, enabling me to identify subtle defects and validate training data against high standards. My advertising studies also help me understand user-centric needs, allowing me to contribute high-quality, contextually relevant data that supports AI model performance.

Labeling Experience

Independent audio producer/musician focused on multilingual voice recording and AI audio quality evaluation (AI tutor-style audio capabilities)

OtherAudioAudio Recording

Self-Operated Independent Audio Producer & Musician producing multilingual voice recordings intended to meet AI training data quality requirements. Used generative audio tools (e.g., Suno AI) to create and iteratively optimize audio content while evaluating output quality and providing actionable feedback for improvement. Performed transcription and prosodic work during voice recording to support clear, natural pronunciation across accents. • Recorded and published Cantonese, Mandarin, and Cantonese-accented Mandarin audio content for training-style datasets • Conducted quality evaluation of AI-generated audio outputs to identify issues and refine results • Performed precise transcription and prosodic annotation/considerations to ensure label-worthy audio quality • Managed end-to-end audio production workflow (recording, editing, mixing, mastering) to deliver consistent training-ready audio

2025 - Present

Education

Work History