For employers

Hire this AI Trainer

Sign in or create an account to invite AI Trainers to your job.

LoginCreate account
Invite to Job
N

Nekay Sandoval

Senior Advanced AI Data Trainer

USA flag
Los Angeles, Usa
$30.00/hrIntermediate

Key Skills

Software

No software listed

Top Subject Matter

AI Training
Safety Domain Expertise
Model Evaluation

Top Data Types

TextText
ImageImage
VideoVideo

Top Task Types

RLHF
Red Teaming
Question Answering
Text Generation
Segmentation
Classification
Object Detection

Freelancer Overview

Senior Advanced AI Data Trainer specializing in model evaluation, RLHF, and safety-focused annotation across text, image, and video modalities. Experienced in assessing model performance through structured rubrics, preference ranking, and adversarial testing to identify failure modes and improve alignment. Background includes red teaming conversational agents for policy violations (e.g., self-harm, sexual content), evaluating LLM outputs for factual accuracy and hallucinations, and ranking responses based on quality and prompt adherence. Additional experience in image generation evaluation (instruction following, visual quality, consistency) and video annotation, including temporal segmentation, content moderation, and engagement-based clip selection. Skilled in applying detailed guidelines and taxonomies to ensure consistent, high-quality labeling across complex tasks. Strong focus on edge cases, ambiguity resolution, and producing reliable training data for safety, ranking, and reasoning benchmarks. Education includes a Bachelor of Science from CSU Dominguez Hills and a UX/UI certification from Avocademy.

IntermediateEnglishSpanish

Labeling Experience

Senior Advanced AI Data Trainer

VideoSegmentation
Selected and annotated high-quality 30-second video clips by identifying engaging segments containing human subjects. Labeled timestamps for clip boundaries and keyframes, including representative frames for individuals. Scored clips based on engagement and relevance to support content ranking and video understanding tasks.

Selected and annotated high-quality 30-second video clips by identifying engaging segments containing human subjects. Labeled timestamps for clip boundaries and keyframes, including representative frames for individuals. Scored clips based on engagement and relevance to support content ranking and video understanding tasks.

2026 - Present

Senior Advanced AI Data Trainer

TextRLHF
As a Senior Advanced AI Data Trainer, I contributed to RLHF-based AI model training, ensuring high standards of safety and effectiveness. I curated and organized diverse text and image data sources for both SFT and diffusion-based image model improvements. My responsibilities included designing evaluation rubrics and leading adversarial testing (red-teaming) to find and resolve model vulnerabilities. • Developed and implemented RLHF protocols and evaluation criteria. • Curated, labeled, and assessed both text and image AI outputs. • Conducted red-teaming to identify and improve adversarial robustness. • Led advanced troubleshooting for accuracy and user experience improvements.

As a Senior Advanced AI Data Trainer, I contributed to RLHF-based AI model training, ensuring high standards of safety and effectiveness. I curated and organized diverse text and image data sources for both SFT and diffusion-based image model improvements. My responsibilities included designing evaluation rubrics and leading adversarial testing (red-teaming) to find and resolve model vulnerabilities. • Developed and implemented RLHF protocols and evaluation criteria. • Curated, labeled, and assessed both text and image AI outputs. • Conducted red-teaming to identify and improve adversarial robustness. • Led advanced troubleshooting for accuracy and user experience improvements.

2023 - Present

Generalist Rubrics Expert

TextRLHF
Designed adversarial and scenario-based prompts to evaluate LLM recommendation capabilities under indirect constraints. Constructed detailed user personas and problem contexts to elicit specific target products without explicitly naming them. Developed structured evaluation rubrics to assess whether model outputs successfully inferred and recommended the intended product based on contextual signals. Calibrated task difficulty by iteratively refining prompts to achieve controlled model failure rates within defined thresholds, enabling robust benchmarking of model performance and sensitivity to implicit user intent.

Designed adversarial and scenario-based prompts to evaluate LLM recommendation capabilities under indirect constraints. Constructed detailed user personas and problem contexts to elicit specific target products without explicitly naming them. Developed structured evaluation rubrics to assess whether model outputs successfully inferred and recommended the intended product based on contextual signals. Calibrated task difficulty by iteratively refining prompts to achieve controlled model failure rates within defined thresholds, enabling robust benchmarking of model performance and sensitivity to implicit user intent.

2025 - 2026

Senior Advance AI Data Trainer

ImageRLHF
Evaluated and ranked AI-generated images using structured rubrics across criteria such as prompt adherence, visual quality, naturalness, and consistency. Contributed to RLHF pipelines by providing human preference data through scoring and pairwise comparisons. Iteratively engineered prompts to test model limits and produce high-quality benchmark outputs.

Evaluated and ranked AI-generated images using structured rubrics across criteria such as prompt adherence, visual quality, naturalness, and consistency. Contributed to RLHF pipelines by providing human preference data through scoring and pairwise comparisons. Iteratively engineered prompts to test model limits and produce high-quality benchmark outputs.

2025 - 2026

Senior Advanced AI Data Trainer

VideoRed Teaming
Annotated video content for sensitive and potentially harmful material, including sexual content, violence, gore, hate speech, and health-related triggers. Applied moderation guidelines to flag violations and context-dependent risks.

Annotated video content for sensitive and potentially harmful material, including sexual content, violence, gore, hate speech, and health-related triggers. Applied moderation guidelines to flag violations and context-dependent risks.

2025 - 2025

Education

A

Avocademy

Certificate, UX/UI Design

Certificate
Not specified
C

CSU Dominguez Hills

Bachelor of Science, Business and Communications

Bachelor of Science
Not specified

Work History

L

Lefty Production Co.

Account & Project Manager

Los Angeles
2018 - 2020
B

BJ Foods Inc.

Executive Assistant

San Diego
2016 - 2018