Looking for a job

Katrina Chan

@katrinachan

AI Evaluation Reviewer at Outlier evaluating Traditional Chinese LLM outputs; sustained a 4.1/5 quality score across 1,380+ tasks.

United States

Message

At Outlier, I evaluate and annotate AI-generated text and conversational outputs across Traditional Chinese projects, sustaining a 4.1/5 quality score over 1,380+ tasks. I identify recurring issues such as mistranslations, hallucinations, and omissions, and write reproducible rationales for my judgments.

In UAT contracts with Hopper, DATAmundi, and Upwork, I tested an AI voice assistant, validated multilingual records, and supported AI training data projects. Earlier, at Vantis Consulting Group, I gathered requirements, supported six UAT phases, and developed ETL processes that improved data management efficiency by 50%.

Experience

Work history, roles, and key accomplishments

Current

AI Evaluation Reviewer

Current

Outlier

Nov 2024 - Present (1 year 10 months)

Promoted to Reviewer across 5+ AI projects, sustaining a 4.1/5 quality score over 1,380+ evaluation tasks. Evaluate and annotate AI-generated text and conversational outputs for linguistic quality, and audit contributor submissions for rubric compliance.

Linguistic Annotation Error Categorization Rubric Based Scoring RLHF Preference Ranking Hallucination Detection Omission Detection Conversational AI Testing Calibration Peer Audit Reproducible Rationales

Current

Freelance QA and Data Specialist

Current

Upwork

Dec 2023 - Present (2 years 9 months)

Completed 32 freelance projects involving UAT/functional testing, data validation, and multilingual voice and text data for AI training.

UAT Functional Testing Data Validation Multilingual Data AI Training Data

Multilingual Data Collection Specialist

DATAmundi

Apr 2026 - Jun 2026 (2 months)

Collected and validated 3,000+ multilingual records (Simplified Chinese, Traditional Chinese, English) against quality standards.

Data Collection Data Validation Multilingual Data Quality Standards

UAT Tester

Hopper

Feb 2026 - May 2026 (3 months)

Placed live test calls to Hopper's AI voice assistant, evaluating full customer-style conversations against expected behaviors across 100+ UAT and regression scenarios.

UAT Regression Testing Conversational AI Testing Intent Detection response evaluation

Business Intelligence Solution Consultant

Vantis Consulting Group

Jul 2021 - Mar 2023 (1 year 8 months)

Client-facing consultant who gathered enterprise requirements, ran 6 UAT phases, and validated outputs before go-live. Authored 3 user manuals, delivered training, and developed ETL processes improving data management efficiency by 50%.