Huang Y.

Key Skills

Software

Other

Top Subject Matter

Robotics Domain Expertise

Embodied AI

Computer Vision

Top Data Types

Image

Geospatial Tiled Imagery

Computer Code Programming

Top Task Types

Object Detection

Classification

Mapping

Segmentation

Point/Key Point

Cuboid

Polygon

Text Generation

Question Answering

Text Summarization

Entity (NER) Classification

Function Calling

Computer Programming/Coding

Data Collection

RLHF

Fine-tuning

Evaluation/Rating

Freelancer Overview

Data Annotator/AI Trainer. Brings 2+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Internal, Proprietary Tooling, and Other. Education includes Bachelor of Science,湖南交通学院and长沙理工大学 (2026). AI-training focus includes data types such as Image, Geospatial, and Tiled Imagery and labeling workflows including Object Detection, Classification, and Mapping.

Labeling Experience

Data Annotator/AI Trainer

ImageObject Detection

At Shenzhen Pudu Robotics, I handled complex multimodal data annotation tasks for service robots, including 2D/3D object detection, semantic segmentation, and behavior annotation. I ensured data compliance and completeness, meticulously following robot perception data standards for model training. I also improved annotation accuracy through rigorous multi-level quality control and team SOP development. • Managed visual, lidar, and action data annotation for robot navigation and embodied intelligence. • Performed and optimized 2D/3D detection, semantic segmentation, keypoint, and behavior labels to meet international acceptance standards. • Developed and automated image and video processing scripts using Python to boost workflow efficiency by 60%. • Led external labeling team management, SOP/training protocol drafting, and collaborated with algorithm team on core AI model updates.

Present

Data Annotator/AI Trainer

ImageClassification

At Changsha Pulan Network Technology, I conducted intensive image and video annotation focusing on image synthesis preferences and 3D model features. I executed annotation classification and property labeling in strict accordance with established project rules to fulfill model training needs. Automation scripts I developed significantly improved annotation speed and reduced error rates for these specialist tasks. • Used project-standard tools to perform classification, recommendation, and property labeling for generated images and videos. • Led script-based batch result validation, format conversion, and summary analytics, lowering error rates to 0.2%. • Co-developed and optimized annotation guides, which increased team annotation accuracy by 5%. • Ensured annotated data matched business and algorithm requirements and drove stable AI model training.

2025 - 2025

GIS Data Annotation Specialist

OtherMapping

In the Amap traffic GIS data labeling project, I was responsible for designing and executing precise annotations of urban road elements on geospatial data. Using GIS tools and Python scripts, I mapped lane types, traffic signs, intersections, and restricted areas, ensuring coordinate accuracy and compliance with mapping requirements. My work supported navigation algorithm optimization and intelligent driving perception model development for the platform. • Performed vector-based geospatial annotation for over 500 km of core district roads. • Created SOPs and optimized labeling rules for challenging road conditions to minimize team error rates. • Led Python script development for GPS correction and data cleansing, improving QC efficiency by 30%. • Achieved a 97.7% annotation precision passing stringent three-step quality control.

2024 - 2024

Data Annotation Project Lead

OtherImageObject Detection

For the AI-powered virtual driving car project, I led efforts in traffic scenario image annotation and dataset curation for training advanced perception models. This involved marking, cleaning, and augmenting traffic images to optimize input data quality for computer vision algorithms. The annotation process directly supported object detection, classification, and model evaluation phases essential for smart driving solutions. • Performed large-scale traffic image labeling for vehicles, obstacles, signs, and pedestrians in diverse scenarios. • Implemented data cleaning and augmentation to boost dataset resilience and model generalization. • Created metrics and tests for model performance, allowing continuous process validation. • Automated the preprocessing and training workflow with Python scripts, cutting training time by more than half.

2023 - 2023

Education

Work History