Adeyemi A.

Key Skills

Software

CVAT

Labelbox

SuperAnnotate

Img Lab

Top Subject Matter

Sport

Ecommerce

Health

Top Data Types

Image

Video

Audio

Top Task Types

Bounding Box

Polygon

Object Detection

Transcription

Segmentation

RLHF

Freelancer Overview

My expertise spans computer vision, healthcare AI, autonomous driving systems, and e-commerce SKU datasets. I handle a wide range of annotation tasks, including bounding boxes, polygon annotation, segmentation, NLP, speech labeling, and quality-focused data analysis. Beyond annotation, I also review, validate, and optimize datasets to minimize errors and reduce costly rework.

Labeling Experience

Image Labelling

ImagePolygon

Worked on image annotation tasks focused on training machine learning models for object detection. The project involved labeling leaves using polygon tools to create accurate training data for AI systems.

2026 - 2026

SPEECH LABELING & SPEAKER ANNOTATION

AudioTranscription

Worked on speech annotation using Annotation Pro, focusing on multi-speaker audio labeling and activity tagging across entire audio files. The task involved listening carefully to audio recordings and accurately identifying speakers based on turn-taking patterns. Each file was fully annotated with speaker labels, starting with speaker_1 and speaker_2, and maintaining consistent labeling whenever a speaker reappeared. A third speaker label (speaker_3) was only introduced when a clearly new speaker was detected beyond the initial two.

2025 - 2025

Data Annotator

ImageBounding Box

I performed detailed annotations to help models learn how to identify and distinguish multiple objects within a single image, including detecting barcodes on product surfaces. The task also required using two segmentation tools within the same asset to ensure precise object boundaries and improve dataset quality.

2025 - 2025

Data Annotator

VideoObject Detection

Worked on a video annotation task focused on object tracking using CVAT. The project involved labeling and tracking multiple objects in football match footage, including footballs, football players, and referees. Each object was assigned a unique ID and consistently tracked across all video frames to ensure accurate temporal annotation. I used closed bounding boxes to fully enclose each object without cutting off any visible part, maintaining precise spatial consistency throughout the sequence.

2025 - 2025

Education

Work History