For employers

Hire this AI Trainer

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

Invite to Job
H

Harry Mwas

AI Specialist – Computer Vision & High-Precision Image Segmentation

ITALY flag
SIENA, Italy
$30.00/hrIntermediateLabelbox

Key Skills

Software

LabelboxLabelbox

Top Subject Matter

Autonomous Systems – Obstacle Detection & Navigation
Robotics – Computer Vision & Spatial Mapping
Geospatial – Satellite Imagery & Land Classification

Top Data Types

ImageImage
VideoVideo

Top Task Types

Bounding Box
Segmentation

Freelancer Overview

With two years of dedicated experience in high-precision image segmentation, I specialize in transforming complex visual data into high-quality training sets for computer vision models. My expertise lies in using Labelbox to execute detailed semantic and instance segmentation across diverse datasets, including [e.g., medical imaging, autonomous driving, or satellite imagery]. I am highly proficient in managing intricate layer hierarchies and ensuring that every pixel is accurately categorized to minimize model error. I pride myself on maintaining a 98%+ quality assurance score, even when handling high-density images with overlapping objects or ambiguous boundaries. I understand the critical role that clean data plays in AI development, and I am adept at following complex ontologies while providing feedback to project managers on edge cases. My goal is always to deliver "gold-standard" data that accelerates model training and deployment.

IntermediateItalianSwahiliEnglish

Labeling Experience

High-Precision Image Segmentation for Computer Vision

ImageSegmentation
Managed end-to-end image segmentation workflows within Labelbox for a large-scale Obstacle Detection project. My primary responsibility involved pixel-level semantic and instance segmentation for a dataset of over 6,500 high-complexity images. I utilized polygons, brushes, and superpixel tools to define precise boundaries for a wide range of obstacles, including pedestrians, cyclists, static debris, and various vehicle classes. I focused on maintaining extreme spatial accuracy, particularly for occluded objects and edge cases in diverse weather conditions. My work directly contributed to improving the model's 'mean Intersection over Union' (mIoU) scores. I consistently exceeded quality benchmarks, maintaining a 98%+ accuracy rating while adhering to a strict class ontology

Managed end-to-end image segmentation workflows within Labelbox for a large-scale Obstacle Detection project. My primary responsibility involved pixel-level semantic and instance segmentation for a dataset of over 6,500 high-complexity images. I utilized polygons, brushes, and superpixel tools to define precise boundaries for a wide range of obstacles, including pedestrians, cyclists, static debris, and various vehicle classes. I focused on maintaining extreme spatial accuracy, particularly for occluded objects and edge cases in diverse weather conditions. My work directly contributed to improving the model's 'mean Intersection over Union' (mIoU) scores. I consistently exceeded quality benchmarks, maintaining a 98%+ accuracy rating while adhering to a strict class ontology

2024 - Present

Education

U

university of siena

masters in economic and data science, economics and data science

masters in economic and data science
2021 - 2023

Work History

F

Freelance AI Specialist

Data Annotation Specialist

SIENA
2024 - Present