Eric P.

Key Skills

Software

Other

Top Subject Matter

LLM output evaluation and prompt design (French/English)

AI red-teaming and bilingual prompt robustness testing

Legal Services & Contract Review

Top Data Types

Text

Image

Document

Top Task Types

Red Teaming

Freelancer Overview

Independent AI Tester & Prompt Designer (Freelance / Personal Projects). Brings 8+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Other. Education includes Secondary 5, Augustin Norbert-Morin School (1999). AI-training focus includes data types such as Text and labeling workflows including Evaluation, Rating, and Red Teaming.

Labeling Experience

Freelance LLM stress-testing / red-teaming (Independent AI Tester & Prompt Designer)

OtherTextRed Teaming

Conducted red-team style evaluations by intentionally crafting unusual questions and bilingual constraints to bypass or expose LLM weaknesses. Tested how models handle French dialogue and edge cases that could trigger language or safety failures. Used findings to adjust prompts and system instructions to steer model behavior toward safer and more helpful responses. • Forced and verified language-handling behavior in bilingual prompts • Probed limitations and safety boundaries using contrived scenarios • Iteratively refined instructions based on observed failures • Used evaluation results to develop stronger prompting intuition

2023 - Present

Independent AI Tester & Prompt Designer (Freelance / Personal Projects)

OtherText

Performed LLM evaluation by testing ChatGPT, Gemini, Perplexity, and other assistants with complex edge-case questions in French and English. Assessed outputs on clarity, factual accuracy, usefulness, tone, and alignment with user intent, then rewrote or suggested improved alternatives when needed. Documented recurring failure modes such as hallucinations, contradictions, and safety issues to inform better prompting. • Designed stress-test prompts to reveal model limitations • Compared assistant responses across multiple models • Refined prompts to improve coherence, quality, and safety • Recorded model failures and identified patterns for improvement

2023 - Present

Education

Work History