Audrey M.

Key Skills

Software

AWS SageMaker

Anno-Mage

Appen

Argilla

Axiom AI

Clickworker

CloudFactory

CrowdFlower

CrowdSource

CVAT

Data Annotation Tech

Dataloop

Datatroniq

Datumbox

Datasaur

Google Cloud Vertex AI

HiveMind

iMerit

Img Lab

LabelImg

Label Studio

Lionbridge

LightTag

Mercor

Micro1

Mighty AI

Mindrift

OpenCV AI Kit (OAK)

Playment

Prodigy

Redbrick AI

Remotasks

Roboflow

Telus

Toloka

Other

Top Subject Matter

Music content annotation (mood, genre, technical attributes) for recommendation and search systems.

Instructional content review and quality evaluation for international audiences.

Large Language Model Alignment & Safety Artificial Intelligence (AI) Generative AI / LLMs Natural Language Processing (NLP) Technology & Cybersecurity

Top Data Types

Audio

Text

Top Task Types

Classification

Freelancer Overview

Music Data Annotator | Beatpulse Labs (AI/data labeling role). Brings 6+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include N and A. Education includes Bachelor of Engineering, Metropolia University of Applied Sciences (2026) and Certificate, Google (2022). AI-training focus includes data types such as Audio and Text and labeling workflows including Classification, Evaluation, and Rating.

Labeling Experience

Music Data Annotator | Beatpulse Labs (AI/data labeling role)

AudioClassification

Music Data Annotator at Beatpulse Labs, performing high-accuracy labeling of audio and associated text-based inputs to support search and recommendation improvements. The work includes applying project-specific guidelines to determine categories tied to audio content attributes and ensuring consistency across entries. Quality control focuses on detecting inconsistencies and low-quality data to protect downstream model training outcomes. • Review and classify audio and text-based data per detailed project guidelines. • Label mood, genre, and technical attributes with a stated goal of 100% classification accuracy. • Identify and flag inconsistencies, ambiguity, or low-quality records. • Collaborate with the technical team to refine annotation guidelines for better model training outcomes.

2026 - Present

Large Language Model Alignment & Safety (Mindrift/Toloka)

TextRLHF

contributed directly to Reinforcement Learning from Human Feedback (RLHF) and Supervised Fine-Tuning (SFT) workflows to evaluate and align leading generative AI models. Evaluated model responses for factual accuracy, logical reasoning, and consistency while detecting and flagging hallucinations, logical fallacies, or policy violations. Processed high volumes of unstructured text datasets under tight turnaround times, strictly applying multi-layered guidelines to rate comparative outputs. Drafted high-quality prompt-response pairs to optimize models for tone, safety, safety guardrails, and conversational helpfulness.

2024 - Present

Online Communication Facilitator | PREC EDU SAIGON (content evaluation)

TextTranscription

Online Communication Facilitator at PREC EDU SAIGON, reviewing and evaluating instructional materials for clarity and accuracy prior to international user delivery. The role required assessing content quality using reasoning to ensure materials were understandable and correct for a cross-cultural audience. The work involved providing detailed documentation and feedback to improve user engagement and platform performance. • Review and evaluate instructional materials for clarity and accuracy. • Provide detailed documentation and feedback on content and user engagement. • Use analytical reasoning to navigate complex cross-cultural instructional challenges. • Support improvements to platform performance based on feedback outcomes.

2022 - 2024

Education

Work History