Pak Tung Chen

Key Skills

Software

Data Annotation Tech

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Image

Text

Top Task Types

Classification

Question Answering

Text Generation

Text Summarization

Translation Localization

Freelancer Overview

An expert in Chatbot training. In the dynamic field of AI, I have rapidly gained valuable expertise in data annotation and AI training data within just two months at Data Annotation Tech. My proficiency lies in English and Chinese bilingual chatbot tasks, where I carefully labeled and refined datasets to enhance the language understanding and response accuracy of chatbots in both Mandarin Chinese & Cantonese Chinese and provide feedbacks/labels in English. This bilingual and cultural capability allows me to bridge language barriers and contribute to the development of more inclusive AI systems.​ Moreover, I possess hands-on experience in data annotation for image/audio chatbots, enabling me to understand the multi-modal nature of AI training. I have also been involved in teaching chatbots to handle unsafe prompts, a crucial aspect of ensuring responsible AI use. This experience showcases my ability to anticipate potential risks (Both general and specific to local culture) and develop strategies to safeguard users according to the rules. My adaptability, attention to detail, and commitment to quality make me a strong asset in shaping reliable and effective AI training data.

Labeling Experience

Bilingual AI Trainer (Data Annotation Tech)

Data Annotation TechTextClassificationQuestion Answering

Text Chatbot Evaluation: Conducted text - only chatbot conversations, compared responses on six axis (relevance, accuracy and etc.), and edited them for better quality.​ Multi - Modal Prompt Work: Created image - integrated prompts, evaluated chatbot responses based on visual context understanding, and refined them as needed.​ Unsafe Prompt Training: Trained AI models to handle unsafe prompts following guidelines, crafting relevant scenarios for effective learning.​ Criteria Development: Formulated fine - grained criteria to steer models towards generating amazing and context-appropriate answers.​ Audio Prompt Analysis: Extracted emotions and cues from audio prompts, guiding models to generate suitable verbal responses.​ General AI - Training: Completed tasks like receipt comparison, enhancing the overall accuracy of AI training data.

Text Chatbot Evaluation: Conducted text - only chatbot conversations, compared responses on six axis (relevance, accuracy and etc.), and edited them for better quality.​ Multi - Modal Prompt Work: Created image - integrated prompts, evaluated chatbot responses based on visual context understanding, and refined them as needed.​ Unsafe Prompt Training: Trained AI models to handle unsafe prompts following guidelines, crafting relevant scenarios for effective learning.​ Criteria Development: Formulated fine - grained criteria to steer models towards generating amazing and context-appropriate answers.​ Audio Prompt Analysis: Extracted emotions and cues from audio prompts, guiding models to generate suitable verbal responses.​ General AI - Training: Completed tasks like receipt comparison, enhancing the overall accuracy of AI training data.

2024

Education

Work History