Dawn G.

Key Skills

Software

Mercor

Other

Encord

Don't disclose

Labelbox

OneForma

Surge AI

Sama

Top Subject Matter

AI safety

trust & safety policy compliance

and high-stakes content moderation for LLM outputs

Top Data Types

Text

3D Sensor

Image

Document

Top Task Types

Classification

Routing

Segmentation

Entity (NER) Classification

Freelancer Overview

Freelance AI Technical Specialist & Content Moderator — Mercor / OneForma / Surge AI. Brings 5+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Mercor, OneForma, and Surge AI. Education includes Bachelor of Science, University of Colorado Boulder (2024) and Google IT Support Professional Certificate, Google. AI-training focus includes data types such as Text, Geospatial, and Tiled Imagery and labeling workflows including Evaluation, Rating, and Entity (NER) Classification.

Labeling Experience

Freelance AI Technical Specialist & Content Moderator - Mercor

LabelboxImageSegmentationClassification

Provided technical evaluation and moderation for AI-generated programming and content in high-stakes environments. Applied structured rubric-based review to assess contextual alignment, accuracy, and policy compliance. Collaborated on QA processes while maintaining strict confidentiality for sensitive projects and data. • Evaluated AI responses for structural appropriateness and programming-logic correctness • Audited dual-response outputs against corporate rubrics for accuracy and empathetic alignment • Moderated and flagged content for adherence to platform and safety policies • Labeled and segmented multimodal data streams (text, image, audio) to support model quality

2024 - Present

Freelance AI Technical Specialist & Content Moderator (Mercor, OneForma, Surge AI)

MercorText

Evaluated and scored AI-generated responses for correctness, contextual alignment, and structural appropriateness in high-stakes programming-related scenarios. Audited dual-response model outputs against rigid corporate rubrics to ensure accuracy, empathetic alignment, and policy compliance. Segmented and labeled multi-modal streams (text, image, audio) to support training of frontier generative models for downstream safety and quality improvements. • Score and refine AI responses using corporate rubrics • Conduct policy compliance checks and risk flagging • Create labeled segments for model training • Maintain QA metrics while ensuring strict confidentiality

2024 - Present

Data Annotator / Associate Candidate (Sama Source)

Sama3D SensorClassification

Executed precise object classification and spatial annotation using image processing and LiDAR software frameworks. Applied logical deduction to resolve nuanced spatial data conflicts and ensure consistent labels across dense batches. Exceeded baseline efficiency and data accuracy KPIs during rapid onboarding and production deployment phases. • Labeled objects in spatial datasets using classification criteria. • Annotated spatial relationships/conflicts for consistency. • Achieved above-baseline accuracy and efficiency targets. • Supported live production annotation workflows.

2023 - 2024

AI Community Rater (TELUS International AI)

Don't discloseTextEntity (NER) Classification

Analyzed live AI-generated search query indices to evaluate factual accuracy and hallucination risk. Assessed alignment with user intent and generated structured technical feedback to improve LLM reasoning. Monitored and flagged policy violations, toxic generation, and high-risk content to enforce safety frameworks. • Evaluated hallucinations and factual accuracy in query outputs. • Rated user intent alignment and reasoning quality. • Produced edge-case documentation for model improvements. • Flagged toxic and policy-violating generations.

2023 - 2024

Peer Support Volunteer (Crisis Operations) — Crisis Text Line / University Wellness Programs

Don't discloseTextRouting

Managed real-time, text-based crisis triage by identifying safety risk signals in incoming messages. Used trauma-informed communication matrices to determine when escalation or routing was required. Routed critical information to appropriate emergency infrastructures while applying linguistic profiling for pacing, tone, and validation cues. • Identified safety risk signals from text conversations. • Determined escalation and routing outcomes using communication matrices. • Routed critical data to emergency workflows. • Applied linguistic profiling for safer conversational handling.

2022 - 2023

Education

Work History