Naomi M.

Key Skills

Software

Data Annotation Tech

Google Cloud Vertex AI

LabelImg

Mercor

Remotasks

Redbrick AI

Roboflow

Top Subject Matter

Finance

E-commerce

C++

Top Data Types

Image

Video

Text

Top Task Types

Point/Key Point

Object Detection

RLHF

Transcription

Evaluation/Rating

Computer Programming/Coding

Data Collection

Prompt + Response Writing (SFT)

Function Calling

Text Summarization

Question Answering

Text Generation

Freelancer Overview

I have a strong background in data annotation, technical translation, and content evaluation, developed through hands-on role experience in AI training data pipelines. My work focuses heavily on evaluating complex datasets, checking linguistic nuances, and optimizing prompt-response pairs to improve the logical reasoning, accuracy, and natural flow of AI models. I am highly comfortable adhering to strict, multi layered style guides, which has allowed mt to consistently deliver high quality data sets that directly improve model performance. What sets me apart is my ability to bridge the gap between technical precision and clear communication. With a background in physics and computer science, I bring a highly analytical mindset to data tagging, making me particularly adept at handling complex technical prompts, algorithmic logic, and structured formatting rules.

Labeling Experience

Data Annotation Specialist

TextRLHF

Project was a large scale optimization pipeline designed to improve a major Large Language Model' reasoning capabilities in STEM and computer science domain, processing hundreds of thousands of prompt-response pairs across a remote team of specialized contributors. On a individual level, I independently evaluated approximately 40 to 60 high complexity technical prompts per week, performing adversarial prompt engineering, response ranking, and detailed text and code editing. My specific tasks involved analyzing AI responses to complex physics theories, mathematics, and C++ algorithms, where I tagged logical or syntax errors, wrote comprehensive justifications for my ratings, and completely rewrote substandard answers to create flawless target data. To ensure the highest data standards, I strictly adhered to a massive, multi layered style guide governing tone and structure, consistently meeting 95% accuracy.

2025 - 2026

Education

Work History