Rocío Gayoso

Key Skills

Software

Appen

CVAT

Labelbox

Label Studio

Mindrift

Prodigy

Toloka

Internal/Proprietary Tooling

Top Subject Matter

Multimodal AI Training and Data Annotation

Top Data Types

Audio

Image

Text

Top Task Types

Audio Recording

Classification

Emotion Recognition

Evaluation Rating

Prompt Response Writing SFT

Freelancer Overview

I'm an AI Trainer and QA Auditor specialized in audio, video and image annotation. I bring a strong professional background in documentary photography, video production, editing, and voice work, supported by over 15 years of hands-on experience with image and audio. I genuinely enjoy working with visual and audio media, which combined with my extensive training and experience, gives me a refined visual and auditory intuition. This allows me to excel in video, image, and audio annotation, as well as dataset creation and management. Detail-oriented, efficient, and highly reliable, I deliver clean, consistent, and high-quality datasets tailored to any project’s needs.

Labeling Experience

Freelance AI Trainer – Freelance (Appen, Outlier, Welocalize)

LabelboxImageClassification

As a Freelance AI Trainer, I specialized in data annotation for text, image, audio, and video projects across multiple platforms and clients. I evaluated outputs from large language models and prompts in both Spanish and English. I collaborated closely with QA teams to ensure guideline adherence and data consistency. • Performed precise tagging, classification, and evaluation of images, audio, video, and text data. • Applied context sensitivity and domain expertise in annotation for multilingual datasets. • Conducted annotation quality assurance and supported guideline development. • Utilized tools including Labelbox, SuperAnnotate, CVAT, Prodigy, LightTag, and Roboflow.

As a Freelance AI Trainer, I specialized in data annotation for text, image, audio, and video projects across multiple platforms and clients. I evaluated outputs from large language models and prompts in both Spanish and English. I collaborated closely with QA teams to ensure guideline adherence and data consistency. • Performed precise tagging, classification, and evaluation of images, audio, video, and text data. • Applied context sensitivity and domain expertise in annotation for multilingual datasets. • Conducted annotation quality assurance and supported guideline development. • Utilized tools including Labelbox, SuperAnnotate, CVAT, Prodigy, LightTag, and Roboflow.

2025 - Present

Wells

Don T DiscloseTextEvaluation Rating

Focused on rating search results on a widely used online music streaming platform, ensuring results matched user intent and relevance guidelines. The labeling involved classifying search results for accuracy and user experience enhancement. This large-scale project involved intensive guideline alignment with regular performance evaluations.

Focused on rating search results on a widely used online music streaming platform, ensuring results matched user intent and relevance guidelines. The labeling involved classifying search results for accuracy and user experience enhancement. This large-scale project involved intensive guideline alignment with regular performance evaluations.

2025

Xilophone Grassland

OtherAudioAudio Recording

Created authentic and natural audio prompts tailored to specific conversation types and categories, to improve AI speech models' natural conversational abilities. Data size was substantial with thousands of audio samples generated, and quality measures included regular peer review and annotation consistency checks.

Created authentic and natural audio prompts tailored to specific conversation types and categories, to improve AI speech models' natural conversational abilities. Data size was substantial with thousands of audio samples generated, and quality measures included regular peer review and annotation consistency checks.

2025 - 2025

Babel

AppenTextEvaluation Rating

The project involved evaluating the quality and safety of LLM-generated prompts and responses translated into Spanish. Tasks included checking translation fluency, grammar, accuracy, and ethical adherence to ensure relevance and correctness in the model’s outputs. Project size was medium, with ongoing collaboration with QA teams to maintain high standards. Strict quality controls including multiple rounds of review and guideline updates were adhered to.

The project involved evaluating the quality and safety of LLM-generated prompts and responses translated into Spanish. Tasks included checking translation fluency, grammar, accuracy, and ethical adherence to ensure relevance and correctness in the model’s outputs. Project size was medium, with ongoing collaboration with QA teams to maintain high standards. Strict quality controls including multiple rounds of review and guideline updates were adhered to.

2025 - 2025

Education

Work History