Dr.michael Kyalo

Key Skills

Software

AWS SageMaker

Anno-Mage

Argilla

Axiom AI

Clickworker

CrowdFlower

Datumbox

Datature

Dataturk

Encord

Mercor

Mindrift

Redbrick AI

Labelbox

Top Subject Matter

No subject matter listed

Top Data Types

Computer Code Programming

Medical Dicom

Text

Top Task Types

Audio Recording

Computer Programming Coding

Cuboid

Data Collection

Evaluation Rating

Freelancer Overview

My journey in data labeling and AI training data has been more than just a career path—it has been an opportunity to contribute to the invisible foundation that makes artificial intelligence systems work. Over the years, I’ve learned that while AI gets the headlines, it’s the careful, detailed work of people like me that makes those systems intelligent, reliable, and useful. Working across text, audio, and image datasets, I’ve built a set of skills that blend precision with problem-solving, patience with speed, and creativity with structure. What excites me most is knowing that the quality of my work directly influences whether a model succeeds or fails, whether it learns the right lessons or gets confused by messy data. That responsibility motivates me to give my best on every project. One of the first things I discovered is that attention to detail is everything. A single mislabeled entry can ripple through a dataset, creating bias or confusing the model. I’ve seen this firsthand during transcription and audio-tagging projects. For example, when background noise in recordings could easily have been mistaken for speech, I resisted the temptation to rush. Instead, I built a system for myself—listening multiple times, cross-checking with context, and flagging unclear cases for review. That carefulness didn’t just protect the quality of the data; it gave me confidence that what I was producing would hold up under real-world testing. These moments taught me that quality control

Labeling Experience

Autonomous Driving Image Annotation and Classification

LabelboxImageBounding BoxPolygon

Project Description I contributed to a large-scale image annotation project designed to train a computer vision model for autonomous driving systems. The project involved labeling tens of thousands of road traffic images under varying conditions such as different lighting, weather, and occlusion. My tasks included drawing bounding boxes around vehicles, pedestrians, and traffic lights, applying polygon segmentation for irregularly shaped objects like road signs, and ensuring classification consistency across multiple categories. The scope of the project demanded exceptional accuracy, as inconsistencies could directly affect the model’s reliability in real-world scenarios. I developed efficient workflows to maintain both speed and precision, batching similar image types, using quality checkpoints, and documenting ambiguous cases for review. I also contributed feedback that helped refine labeling guidelines, which reduced team-wide errors and improved annotation consistency.

Project Description I contributed to a large-scale image annotation project designed to train a computer vision model for autonomous driving systems. The project involved labeling tens of thousands of road traffic images under varying conditions such as different lighting, weather, and occlusion. My tasks included drawing bounding boxes around vehicles, pedestrians, and traffic lights, applying polygon segmentation for irregularly shaped objects like road signs, and ensuring classification consistency across multiple categories. The scope of the project demanded exceptional accuracy, as inconsistencies could directly affect the model’s reliability in real-world scenarios. I developed efficient workflows to maintain both speed and precision, batching similar image types, using quality checkpoints, and documenting ambiguous cases for review. I also contributed feedback that helped refine labeling guidelines, which reduced team-wide errors and improved annotation consistency.

2023 - 2023

Education

Work History