Danny Wachira

Key Skills

Software

AWS SageMaker

Appen

Clickworker

CloudFactory

CVAT

Dataloop

Google Cloud Vertex AI

Labelbox

Label Studio

Lionbridge

OpenCV AI Kit (OAK)

Remotasks

Roboflow

Scale AI

Toloka

Top Subject Matter

self driving, social media posts optimization, fashion industry

llm english and swahili

Satellite image categorization and labeling

Top Data Types

3D Sensor

Image

Video

Top Label Types

Bounding Box

Classification

Cuboid

Polygon

Segmentation

Freelancer Overview

In my extensive experience in data labeling and AI training data, I have demonstrated proficiency in diverse annotation tasks, showcasing a keen eye for detail and a commitment to delivering high-quality labeled datasets. A standout aspect of my skill set lies in lidar annotation and segmentation, where I have successfully applied my expertise to enhance object detection and recognition algorithms in autonomous vehicle systems. My proficiency extends to image annotation and segmentation, where I have tackled complex projects, ensuring accurate and comprehensive labeling for image recognition models. Additionally, I possess a strong background in video annotation, contributing to the development of robust video analysis algorithms. My comprehensive understanding of annotation techniques, coupled with a track record of successful projects, positions me as a valuable asset in the realm of AI training data.

Labeling Experience

Creative writing

Scale AITextPrompt Response Writing SFT

My experience in training large language models (LLMs) and creative writing has been a mix of technical work and creative exploration. On the technical side, I've learned about how these models are built, how to prepare data, and how to fine-tune them to work better. This involves paying close attention to details, learning constantly, and solving problems to make the models better at understanding and creating human-like text. At the same time, my creative writing has helped me with storytelling, developing characters, and exploring themes. This creativity not only fuels my love for writing but also helps me train LLMs, giving them a better sense of language's artistic and expressive side. Together, these experiences combine creativity and technology, driving innovation and deepening my appreciation for the power of language.

My experience in training large language models (LLMs) and creative writing has been a mix of technical work and creative exploration. On the technical side, I've learned about how these models are built, how to prepare data, and how to fine-tune them to work better. This involves paying close attention to details, learning constantly, and solving problems to make the models better at understanding and creating human-like text. At the same time, my creative writing has helped me with storytelling, developing characters, and exploring themes. This creativity not only fuels my love for writing but also helps me train LLMs, giving them a better sense of language's artistic and expressive side. Together, these experiences combine creativity and technology, driving innovation and deepening my appreciation for the power of language.

2022 - 2023

Unnamed Project

CVATVideoBounding Box

This project involved categorizing items within a 200-frame video, each containing lidar points, into various categories. The task involved tracking and appropriately labeling each moving item in the selected video frames. The completion of a single project could span from two to five days.

This project involved categorizing items within a 200-frame video, each containing lidar points, into various categories. The task involved tracking and appropriately labeling each moving item in the selected video frames. The completion of a single project could span from two to five days.

2022 - 2023

House xtp

LabelboxImageSegmentation

This task required annotating an image. A single image was provided, and the objective was to delineate and identify all faults and cracks in properties that needed repair.

This task required annotating an image. A single image was provided, and the objective was to delineate and identify all faults and cracks in properties that needed repair.

2021 - 2022

Whale Semseg

Scale AIImageSegmentation

This task required the annotation of an image, where participants were given a solitary image and tasked with categorizing each pixel into distinct classes like road, building, vehicle, etc. On average, it took approximately 2 hours to complete one assignment.

This task required the annotation of an image, where participants were given a solitary image and tasked with categorizing each pixel into distinct classes like road, building, vehicle, etc. On average, it took approximately 2 hours to complete one assignment.

2020 - 2022

Bee lss

Scale AI3D SensorSegmentation

I categorized point clouds within a 320-frame scene into different classes such as roads, trees, and buildings, etc. The scale of the project was substantial, requiring a minimum of four days for completion. Given the three-dimensional nature of the project, mandatory training sessions were conducted, significantly enhancing my proficiency and accuracy in executing the tasks.

I categorized point clouds within a 320-frame scene into different classes such as roads, trees, and buildings, etc. The scale of the project was substantial, requiring a minimum of four days for completion. Given the three-dimensional nature of the project, mandatory training sessions were conducted, significantly enhancing my proficiency and accuracy in executing the tasks.

2020 - 2022

Education

Work History