Gyula C.

Key Skills

Software

Don't disclose

Top Subject Matter

AI safety evaluation

RLHF/behavior analysis

LLM response quality

Top Data Types

Text

Top Task Types

Evaluation/Rating

RLHF

Red Teaming

Classification

Prompt + Response Writing (SFT)

Question Answering

Freelancer Overview

AI Evaluation & Safety Testing (Independent, 2024–present). Systematic testing and comparative analysis of 7 AI platforms. Tasks: response comparison, safety evaluation, hallucination detection, prompt writing, response rewriting, RLHF documentation, policy-drift detection. Platforms: Pi, Grok, Claude, Gemini, ChatGPT, Copilot, Meta AI. Data type: Text. Workflow: Evaluation, Rating, Red Teaming, Prompt Engineering.

Labeling Experience

AI Evaluation Specialist (Independent / Freelance)

Don't discloseTextEvaluation/Rating

Conducted systematic, safety-focused evaluation of seven major LLM platforms (Pi, Grok, Claude, Gemini, ChatGPT, Copilot, and Meta AI/Muse Spark). Labeled model behaviors by analyzing hallucinations, factual accuracy, policy drift, and consistency across prompts, and documented anomaly patterns. Performed hands-on RLHF-related process documentation and AI behavioral pattern analysis for safety and ethics. • Hallucination detection and factual accuracy verification • Prompt engineering, response evaluation, and response rewriting • Jailbreak/social-engineering recognition and red-teaming style checks • Policy-drift detection and safety anomaly identification

2024 - Present

Education

Work History