Amber P.

Key Skills

Software

Mercor

Labelbox

Other

Top Subject Matter

LLM supervised fine-tuning training data

prompt/response generation

and adversarial evaluation

Top Data Types

Text

Top Task Types

Prompt + Response Writing (SFT)

RLHF

Fine-tuning

Freelancer Overview

AI Training Specialist – Mercor / Alignerr (SFT data creation and rubric-based response writing). Brings 4+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Mercor and Labelbox. Education includes Bachelor of Arts, Georgia Southern University (2010). AI-training focus includes data types such as Text and labeling workflows including Prompt + Response Writing (SFT), RLHF, and Fine-tuning.

Labeling Experience

AI Training Specialist – Mercor / Alignerr (SFT data creation and rubric-based response writing)

MercorTextPrompt + Response Writing (SFT)

Delivered supervised fine-tuning (SFT) work by producing reasoning-heavy training data for frontier LLMs. Designed multi-turn dialogue scaffolds and adversarial prompts to probe logic and chain-of-thought (CoT) behavior. Authored and maintained high-fidelity model responses that satisfy strict technical and rubric-based quality requirements. • Executed multiple high-priority SFT projects concurrently. • Stress-tested model reasoning with adversarial prompts and dialogue structures. • Authored gold-standard responses meeting technical and creative rubrics. • Maintained consistent quality across concurrent gold-standard outputs.

2025 - Present

Advanced AI Writing Specialist – Alignerr

LabelboxTextFine-tuning

Performed expert-level annotation and response writing for high-complexity reasoning tasks used in model fine-tuning workflows. Managed and categorized datasets to support training data preparation in labeling infrastructure. Produced high-fidelity written responses aligned to task requirements and dataset organization needs. • Annotated and authored responses for complex reasoning tasks. • Organized and categorized datasets for fine-tuning. • Utilized Labelbox infrastructure to manage labeling workflows. • Produced training-ready, high-fidelity outputs.

2025 - 2025

AI Content Evaluator — OpenAI Project (via Mercor)

MercorTextRLHF

Contributed to an OpenAI RLHF pipeline through response grading and quality evaluation of model outputs. Identified and mitigated hallucinations using structured feedback and iterative editing loops. Improved factual accuracy, tone consistency, and safety alignment across evaluated responses. • Performed response grading for RLHF quality signals. • Applied structured feedback to address hallucinations. • Iteratively edited outputs to improve evaluation outcomes. • Focused on factual accuracy, tone, and safety alignment.

2025 - 2025

Education

Work History