Edgar Levis

Key Skills

Software

AWS SageMaker

Appen

Argilla

CVAT

Encord

Google Cloud Vertex AI

LabelImg

Label Studio

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Image

Text

Top Task Types

Audio Recording

Computer Programming/Coding

Question Answering

Freelancer Overview

I am an experienced data annotator and AI training specialist with a strong background in labeling text, images, and structured data for machine learning applications. I have worked on tasks such as classification, entity tagging, content categorization, sentiment analysis, and quality assurance of datasets. My approach focuses on accuracy, consistency, and understanding model requirements to produce high-quality training data. I bring strong analytical skills, attention to detail, and hands-on experience using tools like Label Studio, CVAT, and other annotation platforms. I have contributed to projects involving computer vision, NLP, and chatbot training, ensuring that datasets meet the high standards required for AI model performance. My ability to quickly learn new workflows, follow detailed guidelines, and maintain high productivity sets me apart in AI data training projects.

Labeling Experience

Image & Text Annotation for AI Model Training

Label StudioImageClassificationText Generation

I worked on a large-scale data annotation project involving both computer vision and NLP datasets. For the image tasks, I created accurate bounding boxes and object detection labels for products, vehicles, and human activity images. I also performed classification and quality verification to ensure each image met strict annotation standards. On the text side, I completed tasks such as Named Entity Recognition (NER), sentiment classification, intent labeling, and short text generation for chatbot training. This included evaluating LLM responses for correctness, coherence, and safety. The project involved thousands of annotations delivered in batches, requiring high precision, consistent formatting, and adherence to detailed labeling guidelines. I maintained a high-quality score and consistently met project deadlines.

I worked on a large-scale data annotation project involving both computer vision and NLP datasets. For the image tasks, I created accurate bounding boxes and object detection labels for products, vehicles, and human activity images. I also performed classification and quality verification to ensure each image met strict annotation standards. On the text side, I completed tasks such as Named Entity Recognition (NER), sentiment classification, intent labeling, and short text generation for chatbot training. This included evaluating LLM responses for correctness, coherence, and safety. The project involved thousands of annotations delivered in batches, requiring high precision, consistent formatting, and adherence to detailed labeling guidelines. I maintained a high-quality score and consistently met project deadlines.

2024 - 2024

Video Action Recognition & Object Tracking Annotation

AppenVideoPolylineRelationship

I participated in a video annotation project focused on training AI models to recognize human activities, vehicle movements, and object interactions. My tasks included labeling sequences frame-by-frame to track objects across video timelines, identifying specific actions (running, lifting, turning, stopping), and applying polylines for path movement analysis. I also performed temporal segmentation to mark the start and end of actions, ensuring smooth and consistent tracking across long video clips. Quality assurance was a major part of the project, requiring cross-checking labels, maintaining annotation consistency, and following detailed guidelines for acceptable bounding box drift, occlusions, and object disappearance. The project involved hundreds of video clips and required high concentration, strong understanding of motion patterns, and accurate interpretation of action categories. I maintained excellent quality metrics throughout the project duration.

I participated in a video annotation project focused on training AI models to recognize human activities, vehicle movements, and object interactions. My tasks included labeling sequences frame-by-frame to track objects across video timelines, identifying specific actions (running, lifting, turning, stopping), and applying polylines for path movement analysis. I also performed temporal segmentation to mark the start and end of actions, ensuring smooth and consistent tracking across long video clips. Quality assurance was a major part of the project, requiring cross-checking labels, maintaining annotation consistency, and following detailed guidelines for acceptable bounding box drift, occlusions, and object disappearance. The project involved hundreds of video clips and required high concentration, strong understanding of motion patterns, and accurate interpretation of action categories. I maintained excellent quality metrics throughout the project duration.

2023 - 2023

Education

Work History