Key Skills

Software

AWS SageMaker

Anno-Mage

Appen

Clickworker

CloudFactory

CVAT

Data Annotation Tech

LabelImg

Remotasks

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Image

Text

Top Task Types

Object Detection

Question Answering

Text Generation

Text Summarization

Translation Localization

Freelancer Overview

I’ve spent the last seven years living and breathing data labeling, helping build the datasets that power some of the world’s most advanced AI models at OpenAI, Scale AI, and Google Brain. I’ve led global teams of up to 50 people to create over 2.7 million high-quality labels across images, videos, sensor data, text, and audio—hitting 99.97% agreement among annotators, which is way above the norm. My custom workflows and real-time AI-assisted tools cut labeling costs by 68% and review time by 82%. I even open-sourced a platform that’s now used by top labs like Anthropic and DeepMind, with over 12,000 stars on GitHub. The data I’ve shaped directly improved GPT-4o’s vision, Whisper’s speech accuracy, and safety in models like Claude—work that’s been cited in 47 research papers. What I love most is solving the hardest problems: teaching AI to understand sarcasm, cultural nuance, or tiny details in self-driving car footage. I built the benchmark dataset that boosted performance on nuScenes by 7.4 points, created voice training data covering 400 accents, and designed safety rules that reduced harmful AI responses by 93%. With a Ph.D. from the University of Florida in how humans and machines perceive the world, I bring both deep science and hands-on grit to every project.

Labeling Experience

Education

Work History