Oluwakemi Ademuyiwa

Key Skills

Software

Mindrift

Remotasks

Toloka

Scale AI

Other

Top Subject Matter

No subject matter listed

Top Data Types

Audio

Image

Video

Top Task Types

Bounding Box

Polygon

Prompt Response Writing SFT

Segmentation

Text Generation

Freelancer Overview

As a dedicated data annotator with hands-on experience in video labeling and content analysis, I bring a strong foundation in training machine learning models through precise and efficient annotation. At Hugo, I have consistently delivered high-quality data labeling work, ensuring accuracy and adherence to strict project standards. My attention to detail and commitment to maintaining annotation consistency directly contribute to the reliability of AI training datasets. With a background in mass communication and content marketing, I possess sharp analytical skills and an ability to interpret complex guidelines. My experience managing annotation tools, coupled with a proven ability to collaborate remotely, sets me apart as a reliable and adaptable contributor to AI and machine learning projects.

Labeling Experience

Functional description

OtherVideoClassificationQuestion Answering

This project is a comprehensive video annotation task designed to support the training of AI models in human activity recognition. It involves reviewing video footage and accurately labeling segments that depict various daily physical activities such as walking, running, cooking, cleaning, exercising, and other routine movements. Annotators are responsible for identifying activity transitions, segmenting relevant timeframes, and assigning precise labels that reflect the actions being performed. Each annotation must align with predefined activity classes and maintain temporal accuracy to ensure that the AI can learn to detect and distinguish behaviors in real-world scenarios. Quality assurance protocols include consistent cross-checking, validation of activity tags, and adherence to annotation guidelines to ensure reliable, high-quality training data for behavior recognition systems.

This project is a comprehensive video annotation task designed to support the training of AI models in human activity recognition. It involves reviewing video footage and accurately labeling segments that depict various daily physical activities such as walking, running, cooking, cleaning, exercising, and other routine movements. Annotators are responsible for identifying activity transitions, segmenting relevant timeframes, and assigning precise labels that reflect the actions being performed. Each annotation must align with predefined activity classes and maintain temporal accuracy to ensure that the AI can learn to detect and distinguish behaviors in real-world scenarios. Quality assurance protocols include consistent cross-checking, validation of activity tags, and adherence to annotation guidelines to ensure reliable, high-quality training data for behavior recognition systems.

2024

Relations specialist

OtherVideoBounding BoxPolygon

The project is a large-scale video annotation initiative aimed at training AI models to accurately recognize and classify kitchen equipment, tools, and cooking products. It involves drawing precise bounding boxes around these objects across video frames, ensuring each box fits the item tightly with minimal space, allowing for optimal object detection by machine learning systems. Key tasks include frame-by-frame labeling of objects, managing changes in object position, size, and lighting, and adhering to strict annotation guidelines. Quality assurance is a central component, with each annotation reviewed for accuracy, consistency, and compliance with project standards. The scale and complexity of the project require a high level of attention to detail and a strong grasp of annotation tools to produce clean, usable datasets for AI training.

The project is a large-scale video annotation initiative aimed at training AI models to accurately recognize and classify kitchen equipment, tools, and cooking products. It involves drawing precise bounding boxes around these objects across video frames, ensuring each box fits the item tightly with minimal space, allowing for optimal object detection by machine learning systems. Key tasks include frame-by-frame labeling of objects, managing changes in object position, size, and lighting, and adhering to strict annotation guidelines. Quality assurance is a central component, with each annotation reviewed for accuracy, consistency, and compliance with project standards. The scale and complexity of the project require a high level of attention to detail and a strong grasp of annotation tools to produce clean, usable datasets for AI training.

2023 - 2024

Audio Segmentation

TolokaAudioClassificationQuestion Answering

This project focuses on audio segmentation for fashion and lifestyle content, helping to train AI systems to better understand spoken conversations in this field. It involves listening to long audio recordings—like interviews and podcasts—and breaking them into smaller, clear sections based on changes in speaker or topic. Each segment highlights discussions about fashion trends, beauty tips, personal style, and lifestyle topics. To keep the data accurate, each audio clip is double-checked to make sure it starts and ends at the right place, the content flows naturally, and nothing important is cut off. This helps create high-quality data that AI models can learn from easily.

This project focuses on audio segmentation for fashion and lifestyle content, helping to train AI systems to better understand spoken conversations in this field. It involves listening to long audio recordings—like interviews and podcasts—and breaking them into smaller, clear sections based on changes in speaker or topic. Each segment highlights discussions about fashion trends, beauty tips, personal style, and lifestyle topics. To keep the data accurate, each audio clip is double-checked to make sure it starts and ends at the right place, the content flows naturally, and nothing important is cut off. This helps create high-quality data that AI models can learn from easily.

2023 - 2023

Education

Work History