Dexter Garcia

Key Skills

Software

Data Annotation Tech

Labelbox

Mercor

Mindrift

OneForma

Remotasks

Scale AI

SuperAnnotate

Toloka

Telus

Internal/Proprietary Tooling

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Document

Image

Text

Top Label Types

Evaluation Rating

Text Generation

Prompt Response Writing SFT

Classification

Translation Localization

Question Answering

Emotion Recognition

Audio Recording

Transcription

Fine Tuning

RLHF

Freelancer Overview

I am a detail-oriented professional with extensive experience in AI training data, data labeling, and annotation across various platforms. My background includes developing and evaluating prompts, benchmarking LLMs, and enhancing GenAI content quality for educational and STEM-focused projects. I have worked on both English and Filipino (Tagalog) language datasets, including prompt creation, model evaluation, rubric writing, and voice recording for NLP and conversational AI. My technical skills include content development, statistical analysis, LaTeX editing, and quality assurance, and I am adept at collaborating with teams to ensure data accuracy and integrity. With a strong foundation in mathematics education and hands-on experience in AI initiatives, I am committed to delivering high-quality, reliable training data for innovative AI solutions.

Labeling Experience

Prompt and Rubrics/Criteria Writing

Data Annotation TechTextQuestion AnsweringTranslation Localization

- This project consists of different batches with different main goals. - The first few batches centers on writing prompts based on different provided topics and subjects, followed by batches that centers on checking AI's ability to provide correct, acceptable, and safe response, then followed by writing golden criteria that are expected to be followed by the AI bots, and then onwards to more batches of fine-tuning.

- This project consists of different batches with different main goals. - The first few batches centers on writing prompts based on different provided topics and subjects, followed by batches that centers on checking AI's ability to provide correct, acceptable, and safe response, then followed by writing golden criteria that are expected to be followed by the AI bots, and then onwards to more batches of fine-tuning.

2025

Multi-turn conversation with Voice AI

Data Annotation TechAudioFine TuningEvaluation Rating

- In this project, I was tasked to do lots of multi-turn conversations with Voice AI like Gemini and Chat-GPT. - The goal is to evaluate the AI bots ability on different aspects like accuracy, completeness, naturalness, cultural awareness, emotional intent, and others. - Another goal is to evaluate AI-bots ability to transcribe audio in Filipino/Tagalog

- In this project, I was tasked to do lots of multi-turn conversations with Voice AI like Gemini and Chat-GPT. - The goal is to evaluate the AI bots ability on different aspects like accuracy, completeness, naturalness, cultural awareness, emotional intent, and others. - Another goal is to evaluate AI-bots ability to transcribe audio in Filipino/Tagalog

2025

Voice AI

Data Annotation TechAudioPrompt Response Writing SFTAudio Recording

- In this project, I was tasked to write prompts and even do an audio recording for those prompts. - The main goal is to evaluate and train AI bots in transcribing audio.

- In this project, I was tasked to write prompts and even do an audio recording for those prompts. - The main goal is to evaluate and train AI bots in transcribing audio.

2025

Response Rating 3

Scale AITextClassificationEmotion Recognition

This project is composed of different categories, where each category has a particular labeling type. Some of the sections involves: - Evaluating language match of a multi-turn conversation. - Identifying and evaluating emotions in a multi-turn conversation. - Identifying and evaluating tones and insights in a multi-turn conversation. Note: The actual project title is hidden due to confidentiality (NDA)

This project is composed of different categories, where each category has a particular labeling type. Some of the sections involves: - Evaluating language match of a multi-turn conversation. - Identifying and evaluating emotions in a multi-turn conversation. - Identifying and evaluating tones and insights in a multi-turn conversation. Note: The actual project title is hidden due to confidentiality (NDA)

2025 - 2025

Math Project 1 (Golden Response Writing)

LabelboxTextQuestion Answering

The project involves creating an accurate and complete step-by-step solution with final answer to a math question. Note: The actual project title is hidden due to confidentiality (NDA)

The project involves creating an accurate and complete step-by-step solution with final answer to a math question. Note: The actual project title is hidden due to confidentiality (NDA)

2025 - 2025

Show more

Education

Work History