GABRIEL AMARILLE,

Key Skills

Software

Labelbox

Top Subject Matter

No subject matter listed

Top Data Types

Text

Top Label Types

Text Summarization

RLHF

Red Teaming

Prompt Response Writing SFT

Freelancer Overview

I am a senior applied AI scientist with over nine years of experience building production-grade AI systems that rely on high-quality data labeling, annotation, and training data pipelines. My expertise spans neuro-symbolic AI, generative models, multi-modal machine learning, and information retrieval, with hands-on experience designing and deploying NLP, semantic search, document classification, and retrieval-augmented generation (RAG) systems. I have led teams in developing and optimizing data pipelines that integrate structured and unstructured data, ensuring annotated datasets drive model accuracy, fairness, and explainability—especially in regulated domains like legal and finance. My technical toolkit includes Python, TensorFlow, PyTorch, and modern MLOps tools for monitoring and improving data quality and model performance. I am passionate about building robust annotation workflows and evaluation frameworks that support continual learning and adaptive AI, and I thrive in cross-functional teams focused on delivering real-world impact through responsible data-driven AI solutions.

Labeling Experience

GenAI Evaluation & Knowledge-Grounded Data Annotation for Professional Domains

LabelboxTextText SummarizationRLHF

Led and contributed to large-scale data labeling and evaluation initiatives supporting production-grade Generative AI and neuro-symbolic systems in regulated professional domains. The project involved designing and executing high-quality annotation workflows for training, fine-tuning, and evaluating LLM-powered applications used in legal and financial knowledge workflows. Key responsibilities included annotating and reviewing thousands of data samples across multiple task types, including prompt–response pairs for supervised fine-tuning (SFT), human preference ratings for RLHF-style evaluation, entity and concept labeling aligned with domain ontologies, document-level summarization, and question–answer validation grounded in authoritative source documents. Developed detailed annotation guidelines to ensure consistency, correctness, and regulatory alignment, with a strong emphasis on factual accuracy, reasoning quality, explainability, and bias mitigation. Performed adversarial and re

Led and contributed to large-scale data labeling and evaluation initiatives supporting production-grade Generative AI and neuro-symbolic systems in regulated professional domains. The project involved designing and executing high-quality annotation workflows for training, fine-tuning, and evaluating LLM-powered applications used in legal and financial knowledge workflows. Key responsibilities included annotating and reviewing thousands of data samples across multiple task types, including prompt–response pairs for supervised fine-tuning (SFT), human preference ratings for RLHF-style evaluation, entity and concept labeling aligned with domain ontologies, document-level summarization, and question–answer validation grounded in authoritative source documents. Developed detailed annotation guidelines to ensure consistency, correctness, and regulatory alignment, with a strong emphasis on factual accuracy, reasoning quality, explainability, and bias mitigation. Performed adversarial and re

2019

Education

Work History