Xin H.

Key Skills

Software

Other

Top Subject Matter

Mandarin Chinese AI content evaluation

RLHF preference rating

factuality and search quality review

Top Data Types

Text

Document

Top Task Types

RLHF

Freelancer Overview

Independent AI Content & Data Analysis Projects in Contract Review, Compliance, and Legal Research. Brings 2+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Other and Google Sheets. Education includes Football & Market Data Research / Prediction Workflow, Independent Project (2025). AI-training focus includes data types such as Text and labeling workflows including Evaluation, Rating, and RLHF.

Labeling Experience

Spreadsheet-Based Data Annotation & QA (Independent Projects) | Remote | 2025–Present

Text

Undertook spreadsheet-based data annotation and QA on Chinese-language datasets used for analysis and content workflows. Performed data cleaning, consistency checks, and schema validation by finding missing fields, inconsistent labels, and unclear definitions. Ensured dataset reliability by reviewing source-risk issues prior to use in downstream tasks.• Audited dataset fields and labeling consistency in spreadsheets• Performed data cleaning and consistency verification• Checked contradictions and completeness against referenced context• Supported dataset preparation for AI-assisted workflows and reports

2025 - Present

RLHF Preference Rating & Prompt/Response Evaluation (Independent Project) | Remote | 2025–Present

OtherTextRLHF

Reviewed Mandarin Chinese AI outputs to support preference ranking and RLHF-style assessment criteria. Assessed which response is better aligned with user intent and task instructions, using structured acceptance criteria and scoring logic. Performed consistent comparative evaluations to surface contradictions, uncertainty, and instruction deviations for downstream training use.• Rated and compared responses under defined rubrics• Verified instruction adherence and output clarity for Chinese readers• Flagged hallucinations or unsupported claims via evidence comparison• Produced concise summaries of evidence and risk notes to guide selection

2025 - Present

Independent AI Content & Data Analysis Projects (AI Trainer/Chinese Content Evaluator) | Remote | 2025–Present

OtherText

Performed Mandarin Chinese evaluation of AI-generated responses for accuracy, completeness, coherence, tone, usefulness, and intent compliance. Built and applied QA rubrics/checklists and scoring logic to judge whether outputs meet task instructions and factual expectations. Conducted guideline-following reviews including instruction adherence, cultural fit considerations, and risk-aware consistency checks.• Evaluated prompt/response quality and compared candidate outputs• Checked factuality by contrasting evidence from multiple sources• Identified missing fields, unclear definitions, and inconsistent labels in dataset spreadsheets• Documented QA outcomes and maintained separation of facts vs assumptions

2025 - Present

Education

Work History