Yang S.

Key Skills

Software

Scale AI

SuperAnnotate

CrowdSource

Clickworker

Data Annotation Tech

Roboflow

Supervisely

Top Subject Matter

E-commerce – Product Categorization & Customer Support

Finance – Risk Analysis & Fraud Detection

Software Development / AI Tooling

Top Data Types

Text

Image

Document

Top Task Types

Computer Programming/Coding

Text Generation

Text Summarization

Question Answering

Freelancer Overview

Through hands-on AI engineering projects, I have developed a strong data pipeline mindset covering API invocation, output evaluation, and quality assurance. I independently built a DeepSeek API Token Monitoring System that intercepts and analyzes model input/output at millisecond granularity — effectively a real-time "data quality vs. cost" quantification pipeline. I also practice Prompt Engineering extensively, iterating on instruction templates and input formatting to improve model response accuracy and consistency across diverse use cases. In the Codex-Claude Bridge project, I architected an MCP-based dual AI Agent collaboration service responsible for message routing, concurrent request deduplication, and cross-model data passthrough with zero loss and low latency. This required real-time interception, structured processing, and transparent forwarding of model-generated content — equivalent to continuous output quality control and data cleaning in a production AI pipeline. My automation engineering background reinforces a rigorous, script-first approach: repetitive data workflows get automated, ensuring reproducibility and efficient throughput at scale.

Education

Work History