Michael Creates

Key Skills

Software

Mighty AI

Mindrift

Scale AI

Top Subject Matter

No subject matter listed

Top Data Types

Image

Video

Top Label Types

Bounding Box

Polygon

Point Key Point

Polyline

Freelancer Overview

I am an experienced data annotator and labeler with a strong background in managing text, 2D, and 3D image data for AI training, ensuring high accuracy and quality according to client guidelines. My expertise includes image annotation, object detection, semantic and instance segmentation, and text annotation, using tools such as Microsoft Office and Google Workspace. I have a keen eye for detail and a commitment to data privacy, consistently reviewing and correcting annotations for precision and consistency. My experience also extends to digital risk analysis, where I investigated and managed reports on online infringements, demonstrating my analytical skills and ability to work independently under pressure. I am proficient in Python for basic data operations and visualization, and I thrive in fast-paced environments where accuracy and reliability are essential

Labeling Experience

Data Tutor

Mighty AIVideoBounding BoxPolygon

In my last role, I worked on a data annotation project to help train AI for fully autonomous drones. The goal was to make drones smart enough to fly and make decisions on their own in real world situations which involved delivery of packages for clients who purchased them online. My main tasks were labeling images taken from drone cameras. I drew bounding boxes around objects (like people, vehicles, trees, buildings), did pixel-level segmentation to separate things like sky, ground, and obstacles, and added extra information such as distance or whether something was moving. We used tools like segments.ai and labelbox. The project was quite large. My team labeled over 700,000 images in about 35 months. For quality, we followed strict rules. Every image got checked by at least two people, we kept inter annotator agreement above 95%, and the final error rate stayed under 2%. We also had regular reviews, automated checks, and feedback sessions to keep everything consistent and accurate.

In my last role, I worked on a data annotation project to help train AI for fully autonomous drones. The goal was to make drones smart enough to fly and make decisions on their own in real world situations which involved delivery of packages for clients who purchased them online. My main tasks were labeling images taken from drone cameras. I drew bounding boxes around objects (like people, vehicles, trees, buildings), did pixel-level segmentation to separate things like sky, ground, and obstacles, and added extra information such as distance or whether something was moving. We used tools like segments.ai and labelbox. The project was quite large. My team labeled over 700,000 images in about 35 months. For quality, we followed strict rules. Every image got checked by at least two people, we kept inter annotator agreement above 95%, and the final error rate stayed under 2%. We also had regular reviews, automated checks, and feedback sessions to keep everything consistent and accurate.

2021 - 2025

Education

Work History