For employers

Hire this AI Trainer

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

Invite to Job
Y
Youliang

Youliang

Agency
China flagguangdong, China

Key Skills

Software

SuperviselySupervisely
LabelImgLabelImg
RoboflowRoboflow
Data Annotation TechData Annotation Tech
Internal/Proprietary Tooling

Top Subject Matter

No subject matter listed

Top Data Types

ImageImage
TextText
AudioAudio

Top Task Types

SegmentationSegmentation
Text GenerationText Generation
Data CollectionData Collection
Prompt + Response Writing (SFT)Prompt + Response Writing (SFT)

Company Overview

Jiyue Shuke is a China-based AI data collection and annotation service provider specializing in multimodal training data for global AI teams. We support speech, text, image, video, 3D point cloud and GUI screenshot data projects, with strong execution capabilities in Chinese dialect speech, multilingual transcription, SFT/RLHF evaluation, UI task-flow data, autonomous driving, in-cabin interaction, medical, financial and enterprise data scenarios. Our team provides end-to-end services including data sourcing, collection, annotation, quality inspection, anonymization and delivery in structured formats. We support fast pilot runs, flexible scaling, multi-level QA, NDA-based workflows and customized project management. Jiyue Shuke is well suited for AI labs, data companies, model developers, RPA vendors and enterprises seeking reliable China-region data partners for high-quality, compliant and scalable AI training datasets.

Security

Security Overview

Jiyue Shuke operates a comprehensive information security and data privacy governance system in strict compliance with China’s Personal Information Protection Law, Cybersecurity Law, Data Security Law and international data protection norms, covering the full lifecycle of our AI training data services including collection, annotation, quality inspection and delivery. For privacy protection, we enforce standardized de-identification and anonymization procedures across all multimodal data (speech, text, image, video, 3D point cloud, GUI screenshots, etc.). All personally identifiable information (PII) is thoroughly removed before data delivery. All data resources are collected with legitimate authorization and explicit consent of data subjects, and no sensitive personal information is collected beyond the stated project scope. We implement role-based least-privilege access control. All personnel involved in projects must sign strict NDAs and confidentiality agreements before accessing any project data. All data access, operation and transfer behaviors are fully recorded with traceable audit logs. Production data is stored in isolated encrypted environments, and unauthorized downloading, copying or external transmission is strictly prohibited. On infrastructure security, we deploy enterprise-level firewalls, intrusion detection systems and end-to-end encryption for data both in transit and at rest. We conduct regular security vulnerability scans, penetration tests and mandatory security awareness training for all employees. We also follow a standardized data retention and destruction mechanism: data is retained only within the required project cycle, and securely erased upon project completion as contracted. We carry out regular internal security audits and accept client-side compliance inspections to continuously optimize our security and privacy management capabilities, ensuring reliable, compliant and secure data services for global AI teams.

Security Credentials

SOC 2HIPPAISO 27001

Labeling Experience

LabelImg

Full-Stack Multimodal AI Training Data Annotation & RLHF Evaluation Service

LabelImgLabelImgImageImageClassificationClassificationPoint/Key PointPoint/Key Point

This project provides end-to-end AI training data solutions for global AI teams and model developers, covering text, image, audio, video, 3D point cloud and GUI screenshot data types. Our services include SFT instruction annotation, RLHF human preference evaluation, multilingual & Chinese dialect speech transcription, autonomous driving scene labeling, and UI task flow data construction. We enforce multi-level quality inspection, strict data anonymization and NDA-based workflows, with full traceable operation logs. We support fast pilot runs and flexible scaling, and have stably served dozens of model iteration projects with high-quality, compliant training data output.

2023 - 2026