Erastus Kiprop

Key Skills

Software

CVAT

Data Annotation Tech

Diffgram

Labelbox

Roboflow

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Computer Code Programming

Document

Image

Text

Video

Top Label Types

Bounding Box

Segmentation

Classification

Object Detection

Tracking

Freelancer Overview

I am a Computer Science graduate with intermediate-level experience in multimodal AI data annotation and training support across image, video, audio, text, document, and code datasets. I have worked on computer vision projects involving bounding box annotation, object detection, segmentation, YOLO dataset preparation, and frame-by-frame video labeling, including object tracking and action recognition tasks. My experience also includes speech transcription, audio segmentation, speaker identification, sentiment and intent classification, document data extraction, and code review tagging for AI programming models. I am highly detail-oriented and experienced in following strict annotation guidelines while maintaining strong inter-annotator agreement (IAA) and high-quality standards. Beyond labeling, I perform dataset validation, quality assurance checks, and consistency reviews to ensure training data integrity. My technical background in Python, SQL, and backend systems allows me to understand how structured datasets integrate into AI pipelines and model optimization processes. I am reliable, deadline-driven, and committed to delivering precise, scalable, and well-structured training data for AI systems.

Labeling Experience

Video Object Detection and Action Recognition Annotation Project

Data Annotation TechVideoBounding BoxSegmentation

This project focused on video data annotation to support the training of AI models for object detection and action recognition. I worked extensively on labeling video datasets to help improve machine learning systems used in computer vision applications. My role involved carefully reviewing thousands of video frames and drawing bounding boxes around different objects such as people, vehicles, animals, and other environmental elements. I ensured that each object was accurately identified and consistently tracked across consecutive frames. Beyond object detection, I also labeled human activities including walking, running, sitting, driving, and interactions between individuals. I maintained consistent object IDs throughout the videos to support reliable object tracking and temporal analysis. Project scale included: Over 50,000 annotated frames More than 500 labeled video clips Multiple object categories and action classes To maintain high-quality standards, I strictly followed pro

This project focused on video data annotation to support the training of AI models for object detection and action recognition. I worked extensively on labeling video datasets to help improve machine learning systems used in computer vision applications. My role involved carefully reviewing thousands of video frames and drawing bounding boxes around different objects such as people, vehicles, animals, and other environmental elements. I ensured that each object was accurately identified and consistently tracked across consecutive frames. Beyond object detection, I also labeled human activities including walking, running, sitting, driving, and interactions between individuals. I maintained consistent object IDs throughout the videos to support reliable object tracking and temporal analysis. Project scale included: Over 50,000 annotated frames More than 500 labeled video clips Multiple object categories and action classes To maintain high-quality standards, I strictly followed pro

2024 - 2025

Education

Work History