Noor Sagastume

Key Skills

Software

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Top Label Types

RLHF

Fine Tuning

Evaluation Rating

Red Teaming

Prompt Response Writing SFT

Freelancer Overview

I am an AI and NLP specialist with over five years of experience curating, annotating, and validating complex training data for large language models, particularly in STEM domains such as computer science, mathematics, and physics. My work spans prompt engineering, instruction tuning, RLHF evaluation, and rigorous factuality and safety auditing. I have partnered with AI startups and major tech companies to author high-quality instructional data, design red teaming scenarios, and analyze model outputs for accuracy and guideline adherence. Proficient in Python, SQL, and data visualization tools like Tableau, I excel at transforming raw data into actionable insights that directly improve AI model performance, reliability, and safety. My passion lies in shaping AI systems that are more knowledgeable, secure, and helpful.

Labeling Experience

Expert STEM Data Annotation

Scale AIComputer Code ProgrammingRLHFFine Tuning

I serve as a specialized consultant for high-priority AI training projects, focusing on the curation and validation of complex STEM datasets. My work involves authoring high-quality prompt-response pairs for instruction-tuning and performing rigorous RLHF (Reinforcement Learning from Human Feedback) to improve model accuracy and helpfulness. I have curated over 10,000 high-quality data entries, specifically targeting advanced topics in software development, physics, and mathematics. Additionally, I execute 'red teaming' scenarios to identify and mitigate model hallucinations, logical inconsistencies, and safety risks.

I serve as a specialized consultant for high-priority AI training projects, focusing on the curation and validation of complex STEM datasets. My work involves authoring high-quality prompt-response pairs for instruction-tuning and performing rigorous RLHF (Reinforcement Learning from Human Feedback) to improve model accuracy and helpfulness. I have curated over 10,000 high-quality data entries, specifically targeting advanced topics in software development, physics, and mathematics. Additionally, I execute 'red teaming' scenarios to identify and mitigate model hallucinations, logical inconsistencies, and safety risks.

2022 - 2024

Education

Work History