陈平安

Key Skills

Software

Label Studio

Top Subject Matter

Multimodal AI annotation for creative content/branding

Visual content annotation for AI agent development

Top Data Types

Image

Top Task Types

Prompt + Response Writing (SFT)

Freelancer Overview

Visual Content Generation and Annotation Practice Based on Multimodal LLMs – Developer. Core strengths include Label Studio. Education includes Associate Degree, Guangdong Communications Polytechnic (2026). AI-training focus includes data types such as Image and labeling workflows including Prompt + Response Writing (SFT), Evaluation, and Rating.

Labeling Experience

Agent Developer Intensive Practice – Operator

Label StudioImage

As an operator in an intensive agent development practice, I executed over 500 prompt debugging cycles and produced more than 100 high-quality visual data sets. I developed and formalized visual annotation standards by comparing different generative AI model outputs. This role demanded critical evaluation of outputs to establish objective quality criteria for annotations. • Performed extensive prompt debugging targeting edge case scenarios. • Established a repeatable visual annotation standard for benchmark purposes. • Led multi-model evaluation between DALL-E and MLJ outputs. • Ensured consistency and reliability in annotated visual datasets.

2026 - 2026

Visual Content Generation and Annotation Practice Based on Multimodal LLMs – Developer

Label StudioImagePrompt + Response Writing (SFT)

I led the creation and annotation of multi-frame animation content for a skincare brand using multimodal LLMs. My responsibilities included optimizing prompts to improve AI-generated visual outputs and simulating model alignment scoring in training datasets. The project emphasized comparing different model outputs to refine annotation standards. • Utilized advanced prompt engineering to tune visual outputs. • Conducted model comparison between Midjourney and DALL-E for dataset evaluation. • Built annotation workflow integrating creative brand requirements. • Ensured strict adherence to multimodal annotation standards.

2025 - 2026

Education

Work History