Gabriel C.

Key Skills

Software

Labelbox

Scale AI

Prodigy

Don't disclose

Top Subject Matter

Large Language Models (LLMs)

Rlhf Domain Expertise

prompt engineering

Top Data Types

Text

Image

Top Task Types

RLHF

Freelancer Overview

AI Training Specialist / Data Annotator (present). Core strengths include Labelbox, Scale AI, and Prodigy. Education includes Bachelor of Science, University Name. AI-training focus includes data types such as Text and labeling workflows including RLHF, Evaluation, and Rating.

Labeling Experience

AI Training Specialist / Data Annotator (present)

LabelboxTextRLHF

Evaluated and refined LLM-generated outputs using multi-criteria rubrics covering safety, reasoning quality, truthfulness, and style to improve downstream model performance. Designed and ran multi-turn prompt stress tests with adversarial red-teaming to surface edge cases, hallucinations, and security vulnerabilities. Annotated large-scale textual and code-related datasets with advanced alignment scoring schemas aligned to expert benchmarks and used outcomes to support SFT-ready human demonstrations. • Safety and alignment evaluation of model generations • Adversarial red-teaming and edge-case identification • Text/code annotation with alignment scoring against benchmarks • Curation of high-quality human demonstration data for SFT

Present

Content Evaluator / AI Tutor

Don't discloseText

Reviewed and multi-dimensionally graded complex model outputs across technical and domain-specific tasks including code correctness, logical reasoning, and documentation quality. Built gold-standard baseline validation sets to benchmark iterative model checkpoints during production sprints. Identified, categorized, and logged systematic edge-case failures and hallucinations, providing diagnostic data to engineering teams adjusting reward architectures. • Multi-dimensional evaluation and grading of model outputs • Construction of baseline validation/gold-standard datasets • Failure analysis and hallucination cataloging for engineering feedback • High-throughput audit-ready evaluation accuracy in sprints

Not specified

Education

Work History