Osiguru Omolekun

Key Skills

Software

LabelImg

Top Subject Matter

Street Object Detection (Urban Scenes)

Sentiment Analysis (NLP)

Loan Applicant Data (Tabular ML)

Top Data Types

Image

Text

Document

Top Task Types

Bounding Box

Classification

Freelancer Overview

Image Annotation Project Contributor. Core strengths include LabelImg, CVAT, and Google Sheets. Education includes Bachelor of Technology, Federal University of Technology, Akure and West African Senior School Certificate, Christ Ambassadors College, Kaduna (2025). AI-training focus includes data types such as Image, Text, and Document labelling.

Labeling Experience

Data Preprocessing and Encoding Contributor

DocumentClassification

I preprocessed a structured loan applicant dataset with both numerical and categorical fields in preparation for supervised ML model training. Label encoding and one-hot encoding were applied using Scikit-learn's utilities. Normalization and standardization ensured feature consistency and data integrity. • Used Python, Pandas, NumPy, and Scikit-learn for preprocessing and encoding • Focused on preparing tabular data for ML tasks • Included validation of data integrity and consistency checks • Supported downstream predictive model development

I preprocessed a structured loan applicant dataset with both numerical and categorical fields in preparation for supervised ML model training. Label encoding and one-hot encoding were applied using Scikit-learn's utilities. Normalization and standardization ensured feature consistency and data integrity. • Used Python, Pandas, NumPy, and Scikit-learn for preprocessing and encoding • Focused on preparing tabular data for ML tasks • Included validation of data integrity and consistency checks • Supported downstream predictive model development

Not specified

Text Annotation Project Contributor

TextClassification

I labeled over 1,000 tweets and product reviews into positive, negative, and neutral sentiment classes for a natural language processing (NLP) dataset. A systematic labeling rubric was developed and followed, particularly for complex cases like sarcasm and mixed sentiment. Data was cleaned and secondary reviews were undertaken for ambiguous samples. • Used Python, Pandas, and Google Sheets for data annotation and management • Focused on sentiment classification for text data • Emphasized inter-rater consistency and flagging edge cases • Ensured clean, organized, and usable NLP training data

I labeled over 1,000 tweets and product reviews into positive, negative, and neutral sentiment classes for a natural language processing (NLP) dataset. A systematic labeling rubric was developed and followed, particularly for complex cases like sarcasm and mixed sentiment. Data was cleaned and secondary reviews were undertaken for ambiguous samples. • Used Python, Pandas, and Google Sheets for data annotation and management • Focused on sentiment classification for text data • Emphasized inter-rater consistency and flagging edge cases • Ensured clean, organized, and usable NLP training data

Not specified

Image Annotation Project Contributor

LabelimgImageBounding Box

I annotated over 500 street-scene images by drawing bounding boxes around vehicles, pedestrians, traffic lights, and bicycles for an object detection dataset. Strict annotation guidelines were followed to maintain consistency and precision. Quality control reviews were conducted to correct inconsistent or ambiguous labels. • Used LabelImg and CVAT for annotation tasks • Exported in XML (Pascal VOC) and YOLO formats • Focused on detection of multiple street object classes • Ensured the quality and usability of the dataset for model training

I annotated over 500 street-scene images by drawing bounding boxes around vehicles, pedestrians, traffic lights, and bicycles for an object detection dataset. Strict annotation guidelines were followed to maintain consistency and precision. Quality control reviews were conducted to correct inconsistent or ambiguous labels. • Used LabelImg and CVAT for annotation tasks • Exported in XML (Pascal VOC) and YOLO formats • Focused on detection of multiple street object classes • Ensured the quality and usability of the dataset for model training

Not specified

Education

Work History