Bobby Kwong

Key Skills

Software

Scale AI

Telus

Appen

Top Subject Matter

Natural Language Processing

Large Language Models

AI Training

Top Data Types

Text

Image

Audio

Document

Top Task Types

Prompt Response Writing SFT

Classification

Transcription

Freelancer Overview

I have over six years of hands-on experience in AI training data, annotation, and model evaluation across leading platforms such as Outlier AI, Scale AI, TELUS International AI, and Appen. My work has focused heavily on labeling and curating high-quality datasets for machine learning systems, including text, image, video, and speech data. I have contributed to large-scale training pipelines by annotating millions of data points, developing labeling guidelines, and conducting rigorous quality assurance to ensure consistency and accuracy. Additionally, I have extensive experience evaluating AI-generated outputs for relevance, reasoning, and safety, having reviewed over 200,000 model responses and search results. What sets me apart is my deep expertise in prompt engineering and reinforcement learning from human feedback (RLHF), combined with a strong research background (PhD in AI). In my current role, I have created over 4,000 structured prompts and improved model performance by 30% through systematic prompt optimization and feedback loops. I bring a strong understanding of NLP and large language models, along with technical proficiency in Python and ML frameworks, enabling me to bridge the gap between data labeling and model improvement. My ability to design scalable annotation workflows, enhance dataset quality, and deliver consistent, high-accuracy evaluations makes me highly effective in AI training and data annotation roles.

Labeling Experience

Senior AI Trainer / Prompt Engineer

TextPrompt Response Writing SFT

As a Senior AI Trainer and Prompt Engineer at Outlier AI, I designed, optimized, and curated prompts for training large language models. I evaluated AI-generated outputs for accuracy, safety, and contextual quality, providing structured feedback to guide model improvements. I contributed significantly to reinforcement learning workflows and data enhancement strategies. • Developed over 4,000 structured prompts and templates for LLMs. • Evaluated and rated AI responses using established frameworks. • Improved response benchmarks by optimizing prompt strategies. • Supported machine learning teams with actionable labeling feedback.

As a Senior AI Trainer and Prompt Engineer at Outlier AI, I designed, optimized, and curated prompts for training large language models. I evaluated AI-generated outputs for accuracy, safety, and contextual quality, providing structured feedback to guide model improvements. I contributed significantly to reinforcement learning workflows and data enhancement strategies. • Developed over 4,000 structured prompts and templates for LLMs. • Evaluated and rated AI responses using established frameworks. • Improved response benchmarks by optimizing prompt strategies. • Supported machine learning teams with actionable labeling feedback.

2022 - Present

Machine Learning Data Specialist

Scale AIImageClassification

As a Machine Learning Data Specialist at Scale AI, I managed and annotated substantial training datasets for various machine learning and computer vision projects. I performed advanced annotation of image, text, and video data, standardizing guidelines and conducting rigorous QA. I supported engineers by ensuring accuracy and integrity of all labeled data assets. • Annotated millions of data samples for diverse AI systems. • Developed detailed guidelines for high-quality annotation at scale. • Conducted dataset validations and improved data quality metrics. • Enhanced team performance through streamlined labeling processes.

As a Machine Learning Data Specialist at Scale AI, I managed and annotated substantial training datasets for various machine learning and computer vision projects. I performed advanced annotation of image, text, and video data, standardizing guidelines and conducting rigorous QA. I supported engineers by ensuring accuracy and integrity of all labeled data assets. • Annotated millions of data samples for diverse AI systems. • Developed detailed guidelines for high-quality annotation at scale. • Conducted dataset validations and improved data quality metrics. • Enhanced team performance through streamlined labeling processes.

2020 - 2022

AI Data Analyst / Internet Evaluator

TelusText

As an AI Data Analyst and Internet Evaluator at TELUS International AI, I evaluated AI-generated outputs, search results, and recommendation relevance. I assessed the accuracy and intent-alignment of these outputs to improve AI system performance. Additionally, I reviewed web content and search quality metrics using structured evaluation workflows. • Rated and categorized over 200,000 AI outputs for search and recommendation systems. • Conducted structured evaluations for improved search ranking. • Maintained top performance in quality assurance metrics. • Provided detailed feedback to refine AI ranking algorithms.

As an AI Data Analyst and Internet Evaluator at TELUS International AI, I evaluated AI-generated outputs, search results, and recommendation relevance. I assessed the accuracy and intent-alignment of these outputs to improve AI system performance. Additionally, I reviewed web content and search quality metrics using structured evaluation workflows. • Rated and categorized over 200,000 AI outputs for search and recommendation systems. • Conducted structured evaluations for improved search ranking. • Maintained top performance in quality assurance metrics. • Provided detailed feedback to refine AI ranking algorithms.

2019 - 2020

AI Data Annotation Specialist

AppenAudioTranscription

As an AI Data Annotation Specialist at Appen, I annotated and tagged speech, text, and image datasets for various machine learning use cases. My work included participating in speech and voice AI projects and conducting quality checks on all labeled assets. I specialized in natural language dataset annotation for NLP and voice assistant models. • Annotated audio and speech data for conversational AI systems. • Categorized and tagged NLP datasets for model development. • Performed consistent quality assurance on all labeled data. • Contributed to high-impact voice assistant project deployments.

As an AI Data Annotation Specialist at Appen, I annotated and tagged speech, text, and image datasets for various machine learning use cases. My work included participating in speech and voice AI projects and conducting quality checks on all labeled assets. I specialized in natural language dataset annotation for NLP and voice assistant models. • Annotated audio and speech data for conversational AI systems. • Categorized and tagged NLP datasets for model development. • Performed consistent quality assurance on all labeled data. • Contributed to high-impact voice assistant project deployments.

2018 - 2019

AI Microtask Contributor

TextClassification

As an AI Microtask Contributor for Amazon Mechanical Turk, I performed a high volume of human intelligence tasks to support AI and machine learning training pipelines. I specialized in text categorization, dataset tagging, and quality assurance for various AI outputs. My consistent approval rating reflects a rigorous, detail-oriented annotation approach. • Completed over 10,000 microtasks across diverse ML projects. • Validated and tagged datasets for supervised learning. • Performed content classification and reviews for accuracy. • Maintained a 99% approval score on all labeling work.

As an AI Microtask Contributor for Amazon Mechanical Turk, I performed a high volume of human intelligence tasks to support AI and machine learning training pipelines. I specialized in text categorization, dataset tagging, and quality assurance for various AI outputs. My consistent approval rating reflects a rigorous, detail-oriented annotation approach. • Completed over 10,000 microtasks across diverse ML projects. • Validated and tagged datasets for supervised learning. • Performed content classification and reviews for accuracy. • Maintained a 99% approval score on all labeling work.

2017 - 2018

Education

Work History