Funom Magaji

Key Skills

Software

CVAT

Dataloop

Labelbox

Roboflow

SuperAnnotate

Label Studio

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Top Task Types

Classification

Evaluation Rating

Prompt Response Writing SFT

Question Answering

Text Generation

Freelancer Overview

I have hands-on experience in data labeling and AI training across multiple AI domains. I have worked on image annotation tasks, contributing to the preparation of high-quality visual datasets used for computer vision model training. My responsibilities included accurately labeling visual elements to meet strict project guidelines, ensuring consistency, precision, and data quality required for effective machine learning outcomes. In addition, I have participated in search query evaluation for Amazon, where I assessed the relevance between user search queries and product listings to improve e-commerce search accuracy. I also have experience in English large language model (LLM) annotation, evaluating and refining AI-generated text for correctness, relevance, and linguistic quality. Through these projects, I have developed strong attention to detail, guideline adherence, and an understanding of how human feedback improves AI model performance.

Labeling Experience

Search Relevance Task

Label StudioTextEvaluation Rating

I worked on an e-commerce search relevance evaluation project in the Polish language, focused on assessing how well offered products matched user search queries. The task involved manual relevance rating of query–product pairs based on defined relevance categories: Exact, Substitute, Complementary, Irrelevant, and Unjudgeable, following detailed annotation guidelines. Project Size & Quality Measures The project consisted of 100 e-commerce search queries, each evaluated for intent match, product relevance, and overall usefulness to the customer. Quality was maintained through strict guideline adherence, careful linguistic and contextual analysis in Polish, and consistency checks to ensure accurate and reliable relevance judgments across all annotations.

I worked on an e-commerce search relevance evaluation project in the Polish language, focused on assessing how well offered products matched user search queries. The task involved manual relevance rating of query–product pairs based on defined relevance categories: Exact, Substitute, Complementary, Irrelevant, and Unjudgeable, following detailed annotation guidelines. Project Size & Quality Measures The project consisted of 100 e-commerce search queries, each evaluated for intent match, product relevance, and overall usefulness to the customer. Quality was maintained through strict guideline adherence, careful linguistic and contextual analysis in Polish, and consistency checks to ensure accurate and reliable relevance judgments across all annotations.

2025

Education

Work History