Kelvin K.

Key Skills

Software

CVAT

SuperAnnotate

Top Subject Matter

Computer Vision / Image Annotation

NLP / Text Transcription and Labeling

Computer Vision / Video Annotation

Top Data Types

Image

Text

Video

Top Task Types

Object Detection

Transcription

Action Recognition

Freelancer Overview

Freelance Data Labeler / Annotator (Freelance). Core strengths include CVAT, N, and A. Education includes Master of Science, University of Nairobi (2016) and Bachelor of Science, University of Nairobi (2012). AI-training focus includes data types such as Image, Text, and Video and labeling workflows including Object Detection, Transcription, and Action Recognition.

Labeling Experience

Freelance Data Labeler / Annotator (Freelance)

SuperAnnotateVideoAction Recognition

Annotated video data including action-focused segmentation and classification using frame-by-frame labeling workflows. Followed detailed annotation guidelines to ensure temporal consistency and accurate label assignment. Completed quality checks to identify and correct labeling errors before dataset delivery. • Frame-by-frame action labeling. • Video labeling with bounding boxes and segmentation. • Classification labeling aligned to annotation standards. • Quality control to ensure accurate, consistent results.

2016 - Present

Freelance Data Labeler / Annotator (Freelance)

TextTranscription

Performed text transcription and text labeling following verbatim transcription rules and structured formatting requirements. Produced text labels intended for machine learning dataset creation and NLP-related tasks. Maintained high accuracy through careful review and consistency checks. • Verbatim transcription according to provided rules. • Structured text data labeling for dataset readiness. • Consistency-focused annotation to meet guideline expectations. • Review and correction to improve label quality.

2016 - Present

Freelance Data Labeler / Annotator (Freelance)

CVATImageObject Detection

Provided image annotations by drawing bounding boxes and segmenting objects according to detailed guidelines. Ensured label consistency and accuracy through quality checks and corrections. Delivered datasets suitable for downstream machine learning model training and evaluation. • Bounding box creation for objects of interest. • Segmentation of visible object regions. • Compliance with verbatim/structured annotation rules. • Quality assurance to reduce labeling errors.

2016 - Present

Education

Work History