Hank P.

Key Skills

Software

LabelImg

Label Studio

Top Subject Matter

City rubbish detection

Stock price prediction with LSTM and Transformer

Broken logs segmentation

Top Data Types

Computer Code Programming

Image

Top Task Types

Bounding Box

Classification

Object Detection

Segmentation

Freelancer Overview

I am a third-year Artificial Intelligence student at Swinburne University of Technology, Vietnam, with hands-on experience in data labeling and AI model training. My expertise lies in annotating datasets for object detection, particularly using YOLOv11, where I have successfully labeled and managed datasets for real-world applications like trash detection. I am proficient in creating YOLO-compatible datasets, training models, and evaluating their performance using key metrics such as precision, recall, and mAP50. What sets me apart is my practical experience in the full lifecycle of AI training data preparation—from precise labeling to model development and evaluation. My work on a trash detection project demonstrates my ability to deliver accurate and actionable datasets for AI systems. With strong programming skills in Python and attention to detail, I ensure high-quality annotations that contribute to effective AI training.

Labeling Experience

YOLOv11 object detection

Label StudioImageBounding Box

The project aimed to develop an object detection system for classifying and identifying types of trash in urban environments, such as rubbish, furniture, and mattresses. My primary role involved labeling a dataset of 400+ images with high precision, ensuring annotations were YOLOv11-compatible. I used specialized annotation tools and maintained strict adherence to quality standards to ensure consistent and accurate data labeling. To enhance the reliability of the model, I implemented detailed validation processes during data preparation. These included cross-checking annotations and ensuring proper class balance across the dataset. The labeled dataset was then used to train the YOLOv11 model, resulting in a system capable of achieving high precision, recall, and mAP50. This project demonstrated my ability to handle end-to-end data labeling processes while maintaining quality measures critical for AI training.

2024 - 2024