For employers

Hire this AI Trainer

Sign in or create an account to invite AI Trainers to your job.

Invite to Job
Ni Liu

Ni Liu

Generative AI & QA Engineer (Dataset & Annotation)

CHINA flag
Hangzhou, China
$30.00/hrExpertAws SagemakerCVATData Annotation Tech

Key Skills

Software

AWS SageMakerAWS SageMaker
CVATCVAT
Data Annotation TechData Annotation Tech
DataloopDataloop
Surge AISurge AI
SuperviselySupervisely
Google Cloud Vertex AIGoogle Cloud Vertex AI

Top Subject Matter

Generative AI
Computer Vision
Video Model Evaluation

Top Data Types

ImageImage
VideoVideo
TextText
DocumentDocument

Top Label Types

Classification

Freelancer Overview

Generative AI & QA Engineer (Dataset & Annotation). Brings 8+ years of professional experience across legal operations, contract review, compliance, and structured analysis. Core strengths include Internal and Proprietary Tooling. Education includes Master of Science, Zhejiang University (2019) and Bachelor of Science, Jilin University (2016). AI-training focus includes data types such as Image and Video and labeling workflows including Classification, Evaluation, and Rating.

ExpertChinese MandarinEnglish

Labeling Experience

AI Model Evaluator (Video Model Assessment)

Video
I led the evaluation and quality rating of video-based AI models, concentrating on multimodal A/B testing and model selection for tasks like background replacement and gesture recognition. My efforts focused on scrutinizing model outputs for reliability across over 50 configuration variations, utilizing human-in-the-loop processes to ensure optimal selection and robust validation. Automated assessment pipelines and quality gates played a central role in reducing manual QA workload and increasing the consistency of video labeling outcomes. • Executed systematic multimodal video evaluation protocols, including in-depth configuration testing. • Applied HITL reviews to detect failure cases and improve output consistency. • Deployed and refined automated LLM-as-a-Judge workflows for video model assessment in CI/CD environments. • Accelerated model selection by 25% and cut QA effort by over 40% through integrated AI quality controls.

I led the evaluation and quality rating of video-based AI models, concentrating on multimodal A/B testing and model selection for tasks like background replacement and gesture recognition. My efforts focused on scrutinizing model outputs for reliability across over 50 configuration variations, utilizing human-in-the-loop processes to ensure optimal selection and robust validation. Automated assessment pipelines and quality gates played a central role in reducing manual QA workload and increasing the consistency of video labeling outcomes. • Executed systematic multimodal video evaluation protocols, including in-depth configuration testing. • Applied HITL reviews to detect failure cases and improve output consistency. • Deployed and refined automated LLM-as-a-Judge workflows for video model assessment in CI/CD environments. • Accelerated model selection by 25% and cut QA effort by over 40% through integrated AI quality controls.

2025 - Present

Generative AI & QA Engineer (Dataset & Annotation)

ImageClassification
I directly oversaw the creation and management of a training image database, with a primary emphasis on detailed annotation and rigorous data curation workflows. My responsibilities focused on ensuring all image data were accurately labeled to enhance model fine-tuning and visual model performance. Leveraging both manual and automated annotation methods, I implemented strict quality control protocols to maximize annotation reliability and support high-fidelity AI production systems. • Built robust image annotation pipelines and custom curation workflows for enterprise-grade generative AI projects. • Applied ControlNet/GANs for structural anomaly resolution (hands, faces) in outputs, ensuring precision in dataset labeling. • Led HITL (human-in-the-loop) evaluation processes to improve annotation accuracy and production readiness. • Managed the integration of automated and manual annotation stages for efficient, scalable data labeling.

I directly oversaw the creation and management of a training image database, with a primary emphasis on detailed annotation and rigorous data curation workflows. My responsibilities focused on ensuring all image data were accurately labeled to enhance model fine-tuning and visual model performance. Leveraging both manual and automated annotation methods, I implemented strict quality control protocols to maximize annotation reliability and support high-fidelity AI production systems. • Built robust image annotation pipelines and custom curation workflows for enterprise-grade generative AI projects. • Applied ControlNet/GANs for structural anomaly resolution (hands, faces) in outputs, ensuring precision in dataset labeling. • Led HITL (human-in-the-loop) evaluation processes to improve annotation accuracy and production readiness. • Managed the integration of automated and manual annotation stages for efficient, scalable data labeling.

2023 - 2025

Education

Z

Zhejiang University

Master of Science, Electronics and Communication Engineering

Master of Science
2016 - 2019
J

Jilin University

Bachelor of Science, Information Engineering

Bachelor of Science
2012 - 2016

Work History

Z

Zhejiang Insigma International Software

AI Model Evaluator and DevOps Specialist

Hangzhou
2025 - Present
P

PropTexx

Generative AI and QA Engineer

Remote
2023 - 2025