Minyuan W.

Key Skills

Software

No software listed

Top Subject Matter

Construction/real estate architecture AI evaluation and training (LLM/LLM-aligned design assistance)

Code/standards compliance evaluation for architectural outputs

Legal Services & Contract Review

Top Data Types

Text

Document

Top Task Types

Polygon

Freelancer Overview

Architectural Researcher / Project Manager / Solution Expert — Shenzhen Xkool Technology Co., Ltd.. Brings 17+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal and Proprietary Tooling. Education includes Bachelor of Architecture, Beijing Institute of Civil Engineering and Architecture (2010). AI-training focus includes data types such as Text and labeling workflows including RLHF, Evaluation, and Rating.

Labeling Experience

Architectural Researcher / Project Manager / Solution Expert — Shenzhen Xkool Technology Co., Ltd.

TextRLHF

Defined AI product requirements from initial user and domain needs and translated them into functional training and optimization goals. Led live comparisons of AI-generated design outputs and established judgment criteria to support effective human feedback loops. Collaborated with algorithm and engineering teams to iterate strategy and improve output quality, informing what training data should capture and how it should be evaluated. • Created multi-dimensional evaluation dimensions and quality standards for model outputs. • Produced systematic output optimization recommendations based on observed model behavior. • Developed edge-case oriented test prompts to stress instruction-following and compliance. • Acted as a solution expert bridging domain constraints with training/evaluation needs.

2020 - 2025

Chief Architect (Tongyuan Design Group Beijing Branch) — Beijing Yuandao Tonghe Architectural Design Consulting Co., Ltd.

Text

Performed architectural and compliance-focused reviews that supported structured quality assessment of design proposals against explicit rules. Ensured item-by-item verification against national codes and internal standards to flag issues that would degrade downstream AI-aligned outputs. Maintained multi-tier review discipline to generate consistent, actionable feedback for higher-quality deliverables. • Verified compliance against national codes and internal standards. • Flagged issues using structured, item-by-item checks. • Enforced multi-tier review processes to reduce errors. • Translated rule requirements into checkable criteria for evaluation.

2020 - 2020

Education

Work History