Lee Pryce

Key Skills

Software

CloudFactory

CVAT

Data Annotation Tech

Deep Systems

Surge AI

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Text

Video

Top Task Types

Action Recognition

Bounding Box

Computer Programming Coding

Data Collection

Fine Tuning

Freelancer Overview

I have extensive experience in AI training data development and evaluation, specializing in mathematics problem design, solution verification, and data labeling for advanced large language models. At Brilliant.org and through collaborations with UCL, I designed and validated over 400 complex mathematical problems across algebra, probability, number theory, and real-world applications such as cryptography and optimization. Using symbolic algebra tools like SageMath and SymPy, I built systematic solution verification frameworks that improved AI correctness by more than 30%. In addition to content creation, I have worked on dataset curation and quality assurance for AI research projects with platforms such as Scale AI and Outlier, where I developed evaluation rubrics, corrected AI outputs, and ensured alignment with project objectives. My background combines strong subject matter expertise in mathematics with applied AI training skills, enabling me to create high-quality, contextually relevant datasets that directly enhance AI reasoning, accuracy, and scalability.

Labeling Experience

Mathematics Problem Design & AI Data Labeling for LLM Training

Surge AITextQuestion AnsweringText Generation

Developed and labeled large-scale datasets for training and evaluating large language models in advanced mathematics. Tasks included generating over 400 original problems in algebra, probability, and number theory, writing step-by-step solutions, and creating evaluation rubrics to measure AI performance. Labeled AI outputs for correctness, logical consistency, and clarity while applying automated proof-checking tools (SageMath, SymPy). Collaborated with AI researchers to align datasets with project objectives, improving mathematical reasoning accuracy of LLMs by 30%. Ensured dataset quality through systematic review cycles and statistical validation methods.

Developed and labeled large-scale datasets for training and evaluating large language models in advanced mathematics. Tasks included generating over 400 original problems in algebra, probability, and number theory, writing step-by-step solutions, and creating evaluation rubrics to measure AI performance. Labeled AI outputs for correctness, logical consistency, and clarity while applying automated proof-checking tools (SageMath, SymPy). Collaborated with AI researchers to align datasets with project objectives, improving mathematical reasoning accuracy of LLMs by 30%. Ensured dataset quality through systematic review cycles and statistical validation methods.

2021 - 2023

Education

Work History