Tian M.

Key Skills

Software

Don't disclose

Label Studio

Top Subject Matter

LLM-powered enterprise knowledge Q&A

Graphrag Domain Expertise

and evaluation

Top Data Types

Text

Audio

Top Task Types

Fine-tuning

Freelancer Overview

AI Application Engineer / RAG (WiseChain Technology (Shenzhen) Co., Ltd.). Brings 5+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Don't disclose and Label Studio. Education includes Bachelor of Science, N/A (2021). AI-training focus includes data types such as Text and labeling workflows including Evaluation, Rating, and Fine-tuning.

Labeling Experience

Intelligent Service Agent System (labeling platform and dataset standardization)

Label StudioTextFine-tuning

Developed an intelligent labeling platform using LLM-assisted labeling and verification workflows for dataset creation. Standardized datasets and distributed labeling tasks across multiple users to improve consistency. Integrated labeling outputs into the enterprise GraphRAG pipeline to support downstream retrieval and response performance. • Used Label Studio and LLM to run labeling and annotating verification • Integrated Label Studio + LLM to standardize datasets and coordinate multi-user task distribution • Built schema discovery units (topics/facts) and standardized labels, tags, and nodes for GraphRAG • Optimized labeled data utilization for entity and graph-based retrieval quality

2025 - Present

AI Application Engineer / RAG (WiseChain Technology (Shenzhen) Co., Ltd.)

Don't discloseText

Built and optimized intelligent Q&A and retrieval pipelines that required evaluation and tuning of model outputs for improved performance. Implemented RAG approaches and test workflows to measure effectiveness across traditional FAQ and GraphRAG scenarios. Worked on reducing latency in query planning and execution using iterative evaluation during online testing and deployment. • Increased performance score from 0.82 to 0.93 for FAQ scenarios through evaluation-driven improvements • Improved Neo4j + Text2Cypher + GraphRAG efficiency and reduced planner/map-reduce execution time from 4.8s to 2.1s • Supported multilingual enterprise Q&A by iterating retrieval and response effectiveness • Conducted continuous system testing to validate reasoning and retrieval quality

2025 - Present

Education

Work History