Festus Kunmi Ominiyi

Key Skills

Software

Appen

CrowdSource

Label Studio

Lionbridge

Telus

CVAT

LabelImg

Top Subject Matter

Speech Recognition

Autonomous Vehicles

Medical NER

Top Data Types

Audio

Document

Image

Text

Video

Top Label Types

Action Recognition

Audio Recording

Bounding Box

Data Collection

Transcription

Translation Localization

Entity Ner Classification

Freelancer Overview

Data Annotator (Audio Transcription and Speaker Diarization). Core strengths include Lionbridge, CVAT, and LabelImg. Education includes Bachelor of Science, Ladoke Akintola University Of Technology (2019). AI-training focus includes data types such as Audio, Image, and Text and labeling workflows including Transcription, Bounding Box, and Entity (NER) Classification.

Labeling Experience

Data annotation

LionbridgeAudioBounding BoxData Collection

Labeling: Tagging objects in images, transcribing speech to text, annotating entities in documents, or classifying sentiment. Quality Assurance: Reviewing annotations for accuracy and consistency using validation checks and inter-annotator agreement. Guideline Development: Creating clear annotation rules and instructions for human annotators to ensure uniformity. Tooling & Workflow: Using specialized software for efficient labeling, task assignment, and progress tracking. Common Applications: Used in computer vision (e.g., self-driving cars, medical imaging), natural language processing (e.g., chatbots, translation), speech recognition, and many other AI-driven systems. Goal: To produce clean, structured, and reliable datasets that enable machine learning models to be trained effectively and perform accurately in real-world applications.

Labeling: Tagging objects in images, transcribing speech to text, annotating entities in documents, or classifying sentiment. Quality Assurance: Reviewing annotations for accuracy and consistency using validation checks and inter-annotator agreement. Guideline Development: Creating clear annotation rules and instructions for human annotators to ensure uniformity. Tooling & Workflow: Using specialized software for efficient labeling, task assignment, and progress tracking. Common Applications: Used in computer vision (e.g., self-driving cars, medical imaging), natural language processing (e.g., chatbots, translation), speech recognition, and many other AI-driven systems. Goal: To produce clean, structured, and reliable datasets that enable machine learning models to be trained effectively and perform accurately in real-world applications.

2024

Audio Transcription & Speaker Tagging (Telus Digital International)

AudioTranscription

Transcribed and labeled conversational audio data with timestamps and speaker IDs for a voice assistant dataset. Ensured the accuracy and clarity of transcriptions to assist AI model development. Handled diverse dialogue scenarios to improve model robustness. • Labeled conversation turns and speaker changes precisely. • Maintained comprehensive timestamps throughout recordings. • Designed to mimic real-world assistant interaction scenarios. • Enhanced training quality for mock voice assistant AI.

Transcribed and labeled conversational audio data with timestamps and speaker IDs for a voice assistant dataset. Ensured the accuracy and clarity of transcriptions to assist AI model development. Handled diverse dialogue scenarios to improve model robustness. • Labeled conversation turns and speaker changes precisely. • Maintained comprehensive timestamps throughout recordings. • Designed to mimic real-world assistant interaction scenarios. • Enhanced training quality for mock voice assistant AI.

Not specified

Image Annotation for Object Detection (Telus Digital International)

CVATImageBounding Box

Annotated images using bounding boxes for household item recognition in a personal computer vision project. Utilized industry-standard annotation platforms for detailed image labeling. Refined annotations to ensure accurate model training outcomes. • Used CVAT and LabelImg to annotate object classes. • Labeled an unspecified number of images for model evaluation. • Focused on household item categories and object boundaries. • Improved dataset quality through careful verification.

Annotated images using bounding boxes for household item recognition in a personal computer vision project. Utilized industry-standard annotation platforms for detailed image labeling. Refined annotations to ensure accurate model training outcomes. • Used CVAT and LabelImg to annotate object classes. • Labeled an unspecified number of images for model evaluation. • Focused on household item categories and object boundaries. • Improved dataset quality through careful verification.

Not specified

Data Annotator (Medical Text NER)

LionbridgeTextEntity Ner Classification

Performed Named Entity Recognition (NER) on medical text for a machine learning training project. Consistently achieved 99% accuracy in identifying and tagging key entities. Adhered rigorously to strict annotation guidelines set by project leads. • Tagged medical terms and entities in expansive datasets. • Ensured compliance with data privacy and domain standards. • Adjusted NER boundaries based on feedback from ML engineers. • Maintained high annotation quality through iterative reviews.

Performed Named Entity Recognition (NER) on medical text for a machine learning training project. Consistently achieved 99% accuracy in identifying and tagging key entities. Adhered rigorously to strict annotation guidelines set by project leads. • Tagged medical terms and entities in expansive datasets. • Ensured compliance with data privacy and domain standards. • Adjusted NER boundaries based on feedback from ML engineers. • Maintained high annotation quality through iterative reviews.

Not specified

Data Annotator (Image Annotation for Object Detection)

LionbridgeImageBounding Box

Labeled images with bounding boxes and segmentation masks for object detection in autonomous vehicle datasets. Ensured high data quality for AI/ML training purposes. Closely followed project-specific guidelines to deliver consistent annotations. • Labeled 1,500+ images for object detection tasks. • Applied both bounding box and segmentation techniques. • Worked under tight deadlines to meet project goals. • Supported the development of autonomous vehicle perception models.

Labeled images with bounding boxes and segmentation masks for object detection in autonomous vehicle datasets. Ensured high data quality for AI/ML training purposes. Closely followed project-specific guidelines to deliver consistent annotations. • Labeled 1,500+ images for object detection tasks. • Applied both bounding box and segmentation techniques. • Worked under tight deadlines to meet project goals. • Supported the development of autonomous vehicle perception models.

Not specified

Show more

Education

Work History