Hamza Lagrari

Key Skills

Software

Labelbox

Label Studio

Roboflow

Top Subject Matter

Ai agent

blockchain

cars , soccer

Top Data Types

Text

Image

Top Label Types

Prompt Response Writing SFT

Polygon

Freelancer Overview

I am a data annotation and AI evaluation specialist with over two years of hands-on experience in data labeling for machine learning, particularly within NLP projects. My background includes evaluating AI-generated responses for accuracy, relevance, and safety, as well as performing detailed annotation work following strict guidelines and quality standards. I am proficient with tools such as Labelbox, Label Studio, and Roboflow, and I have a strong foundation in data formats like JSON and YAML. My technical skills are complemented by a solid understanding of software testing and programming, allowing me to ensure the quality and reliability of training data used in AI and NLP workflows.

Labeling Experience

Multi-Turn AI Chat | Rating Task (Arabic-Moroccan)

LabelboxTextPrompt Response Writing SFT

In this role, I was responsible for reviewing multi-turn AI chat conversations, carefully reading all exchanges between the user and the AI to evaluate the quality of responses. At the end of each session, I analyzed two AI-generated answers, rated them based on predefined quality metrics such as coherence, accuracy, relevance, and fluency, and selected the response that best fit the context. In addition, I refined low-quality responses when necessary to improve conversational flow and engagement. My contributions helped enhance AI language models, making them more natural, accurate, and effective for real-world interactions

In this role, I was responsible for reviewing multi-turn AI chat conversations, carefully reading all exchanges between the user and the AI to evaluate the quality of responses. At the end of each session, I analyzed two AI-generated answers, rated them based on predefined quality metrics such as coherence, accuracy, relevance, and fluency, and selected the response that best fit the context. In addition, I refined low-quality responses when necessary to improve conversational flow and engagement. My contributions helped enhance AI language models, making them more natural, accurate, and effective for real-world interactions

2025 - 2025

Architectural Floor Plan Segmentation & Room-Type Annotation Project

RoboflowImagePolygon

I worked on a computer vision dataset focused on analyzing and segmenting architectural floor plan images. The goal was to create high-quality training data for an AI model that detects and classifies room types in floor plans.

I worked on a computer vision dataset focused on analyzing and segmenting architectural floor plan images. The goal was to create high-quality training data for an AI model that detects and classifies room types in floor plans.

2024 - 2024

Click worker

Label StudioImagePolygon

I worked on image labeling tasks where I was responsible for identifying and tagging specific objects within images based on detailed project guidelines. This involved carefully analyzing visual data, classifying objects accurately, and ensuring high-quality annotations to support AI model training. My work contributed to improving computer vision systems by providing precise and consistent labeled datasets used for machine learning applications

I worked on image labeling tasks where I was responsible for identifying and tagging specific objects within images based on detailed project guidelines. This involved carefully analyzing visual data, classifying objects accurately, and ensuring high-quality annotations to support AI model training. My work contributed to improving computer vision systems by providing precise and consistent labeled datasets used for machine learning applications

2024 - 2024

Education

Work History