Pierre-alex Guinan

Key Skills

Software

Appen

Data Annotation Tech

Mercor

Micro1

Top Subject Matter

Text Evaluation

Prompt Engineering

Harm Taxonomy Classification

Top Data Types

Image

Text

Video

Document

Top Task Types

Classification

Evaluation Rating

Prompt Response Writing SFT

RLHF

Translation Localization

Data Collection

Freelancer Overview

Bilingual AI training specialist working across four concurrent platforms in data annotation, model evaluation, and Reinforcement Learning from Human Feedback (RLHF) data production. Core work includes safety-critical Natural Language Processing (NLP) evaluation in French, adversarial prompt engineering, rubric-based quality scoring, and egocentric video capture for robotic learning. Native French, advanced English. BBA in progress at HEC Montréal.

Labeling Experience

Heracles: Basketball Expert

TextEvaluation Rating

Score AI-generated basketball analysis for accuracy and sports-specific terminology.

Score AI-generated basketball analysis for accuracy and sports-specific terminology.

2026 - Present

Heracles: EU Football/Soccer Expert

TextEvaluation Rating

Evaluate AI-generated EU football commentary for factual precision, tonal quality, and domain-specific knowledge.

Evaluate AI-generated EU football commentary for factual precision, tonal quality, and domain-specific knowledge.

2026 - Present

Kardia Halo: Health Image Sourcing and Localization

ImageEvaluation Rating

Localized health-domain image prompts into French (Canada) and sourced regionally appropriate visual assets matching prompt intent and cultural context.

Localized health-domain image prompts into French (Canada) and sourced regionally appropriate visual assets matching prompt intent and cultural context.

2025 - Present

Merlin – Safety Conversation Labeling

Data Annotation TechTextEvaluation Rating

Labeled French AI conversations for safety concerns, including harassment, bias, and unsafe advice. Applied harm taxonomies and documented rationales to support AI safety research. Ensured compliance with ethical annotation standards while improving model safety and contextual understanding.

Labeled French AI conversations for safety concerns, including harassment, bias, and unsafe advice. Applied harm taxonomies and documented rationales to support AI safety research. Ensured compliance with ethical annotation standards while improving model safety and contextual understanding.

2025 - Present

Selenite – Rubric Clean-Up

Data Annotation TechTextEvaluation Rating

Reviewed and standardized rubric structures to reduce ambiguity and improve inter-annotator reliability. Refined evaluation criteria targeting subtle unsafe behaviors such as bias, misinformation, and manipulative tone. Ensured coherence and precision in large-scale model evaluation frameworks.

Reviewed and standardized rubric structures to reduce ambiguity and improve inter-annotator reliability. Refined evaluation criteria targeting subtle unsafe behaviors such as bias, misinformation, and manipulative tone. Ensured coherence and precision in large-scale model evaluation frameworks.

2025 - Present

Show more

Education

Work History