Adedamola A.

Key Skills

Software

AWS SageMaker

LabelImg

Roboflow

Top Subject Matter

Generative AI & Large Language Models (LLMs)

Autonomous Vehicles & Advanced Robotics

E-commerce & Retail AI

Top Data Types

Text

Image

Top Task Types

Text Summarization

Classification

Diagnosis

Segmentation

Object Detection

Transcription

Data Collection

Freelancer Overview

As an ML Engineer, my biggest advantage in data annotation is that I actually build and deploy the models that rely on this data. I have practical experience labeling computer vision datasets, such as a Waste Disposal Object Detection project where I categorized images into paper, glass, carton, nylon, plastic, and others. Because I handle projects from start to finish, I understand exactly how bad data can ruin a model's performance. My experience in fast-paced competitive data science also means I learn very fast. I can quickly adapt to new tools or complex guidelines and deliver top-quality work without needing someone to hold my hand.

Labeling Experience

Data Engineering Intern (LLM-integrated summarization/QA/entity extraction and sentiment labeling/enrichment) at Data Epic Remote

TextText Summarization

Created a Python CLI workflow that uses the Gemini API to generate text summaries, context-aware question answering, and structured entity extraction from input text. Built an end-to-end pipeline that enriches e-commerce review content by running LLM-based sentiment classification and syncing the resulting outputs to a reporting dashboard. Ensured reliability by adding robust external API handling including retries, custom exceptions, and rate-limit management. • Implemented Gemini API calls for summarization, QA, and entity extraction. • Developed LLM-driven sentiment classification for review enrichment. • Managed output synchronization to a Google Sheets dashboard. • Added fault-tolerant GitHub API client logic for safe data retrieval.

2025 - 2025

Machine Learning Intern at Plethudeep Hub (classification pipeline development for labeled prediction tasks)

Classification

Built machine-learning classification pipelines for loan approval and customer churn, including feature engineering and model serving as containerized REST APIs. While not explicitly described as manual annotation, the work involved constructing labeled targets (e.g., churn labels) and generating predictions for downstream inference. Delivered production-ready inference endpoints for real-time and batch scoring. • Implemented a loan approval classification model deployed via Flask API in Docker. • Built an end-to-end churn prediction pipeline from feature engineering to serving. • Containerized ML inference using Dockerized Flask endpoints. • Produced predictive outputs for real-time application use.

2024 - 2024

Customer Retention Intelligence project (label engineering and predictive scoring via deployed API)

TextClassification

Created churn/retention intelligence analytics by engineering churn labels from raw transactional retail data using a defined observation/performance methodology. Trained and served ML models (segmentation and classification) that output churn scores for downstream consumption through a REST API. Packaged the system for real-time scoring with batch inference support. • Engineered churn labels from scratch using observation/performance windows. • Built K-Means segmentation and trained a Random Forest churn classifier on RFM features. • Containerized and deployed a FastAPI REST service for real-time churn scoring. • Supported batch inference for larger offline runs.

Not specified

Oral Disease Detection project (AI training/inference for diagnostic classification)

Diagnosis

Developed a histopathology image screening application for oral disease detection by training and evaluating a deep learning model for diagnostic classification. Fine-tuned a ResNet50 model using PyTorch to produce predictions from medical images for clinical early screening workflows. Packaged the trained model behind a FastAPI backend for production inference. • Fine-tuned ResNet50 on histopathological images for oral disease classes (OSCC/dysplasia). • Achieved high test accuracy and low inference latency for clinical screening use. • Containerized the application with Docker for deployment readiness. • Served model predictions via FastAPI backend.

Not specified

Education

Work History