Ahmed Hisham

Key Skills

Software

Labelbox

Data Annotation Tech

Top Subject Matter

No subject matter listed

Top Data Types

Text

Top Task Types

Text Generation

Evaluation Rating

Prompt Response Writing SFT

Freelancer Overview

I am a native Arabic speaker with hands-on experience in data annotation, AI training, and NLP evaluation, specializing in Arabic language tasks. My work has focused on intent detection, semantic similarity, and chatbot evaluation for projects such as Argon and Omine, where I evaluated and edited AI responses for fluency, cultural relevance, and safety. I am skilled in using annotation platforms like CrowdGen and DA Tools, and have a strong technical background in Python, Java, and data analysis tools such as Pandas and SQL. My academic background in computer science, combined with practical experience in model behavior analysis and ambiguity resolution, enables me to deliver high-quality linguistic input that improves AI model performance. I am passionate about bridging the gap between language and technology, and excel at collaborating with teams to enhance the accuracy and naturalness of AI systems.

Labeling Experience

Omni

LabelboxTextText GenerationEvaluation Rating

I contributed to Project Omni, a data labeling initiative focused on evaluating conversational AI responses to enhance natural language processing (NLP) models. My role involved analyzing user prompts and comparing two AI-generated responses (A and B) using the Labelbox platform. For example, I evaluated responses to a prompt about challenges in creating fusion cuisine, ensuring alignment with the AI persona (a chef specialized in Indian fusion cuisine). I tagged prompts, selected the better response based on relevance and accuracy, and rated both responses for language quality, including grammar, fluency, and consistency, achieving a 95% accuracy rate in my annotations. This project honed my attention to detail and deepened my understanding of conversational AI evaluation, directly supporting the improvement of AI-driven dialogue systems.

I contributed to Project Omni, a data labeling initiative focused on evaluating conversational AI responses to enhance natural language processing (NLP) models. My role involved analyzing user prompts and comparing two AI-generated responses (A and B) using the Labelbox platform. For example, I evaluated responses to a prompt about challenges in creating fusion cuisine, ensuring alignment with the AI persona (a chef specialized in Indian fusion cuisine). I tagged prompts, selected the better response based on relevance and accuracy, and rated both responses for language quality, including grammar, fluency, and consistency, achieving a 95% accuracy rate in my annotations. This project honed my attention to detail and deepened my understanding of conversational AI evaluation, directly supporting the improvement of AI-driven dialogue systems.

2024

Argon

Data Annotation TechTextText GenerationEvaluation Rating

I participated in the Argon project, which aimed to enhance a conversational AI model by evaluating and refining chatbot responses in Egyptian Arabic. My responsibilities included crafting prompts in Egyptian Arabic, comparing responses from two models (A and B) across five quality dimensions, and editing responses to improve localization, grammar, and conciseness. Using a custom platform, I ensured responses aligned with Egyptian Arabic conventions, correcting spelling, word choice, and cultural nuances. I evaluated and edited hundreds of responses, maintaining a 97% accuracy rate in my annotations and edits, contributing to a more natural and culturally relevant chatbot experience. Additional Information (Optional): The project required fluency in Egyptian Arabic and strict adherence to localization guidelines, which sharpened my skills in linguistic precision and cultural adaptation. I frequently incorporated system prompts (e.g., “You are a friendly chatbot using Egyptian slang”) t

I participated in the Argon project, which aimed to enhance a conversational AI model by evaluating and refining chatbot responses in Egyptian Arabic. My responsibilities included crafting prompts in Egyptian Arabic, comparing responses from two models (A and B) across five quality dimensions, and editing responses to improve localization, grammar, and conciseness. Using a custom platform, I ensured responses aligned with Egyptian Arabic conventions, correcting spelling, word choice, and cultural nuances. I evaluated and edited hundreds of responses, maintaining a 97% accuracy rate in my annotations and edits, contributing to a more natural and culturally relevant chatbot experience. Additional Information (Optional): The project required fluency in Egyptian Arabic and strict adherence to localization guidelines, which sharpened my skills in linguistic precision and cultural adaptation. I frequently incorporated system prompts (e.g., “You are a friendly chatbot using Egyptian slang”) t

2024 - 2024

Education

Work History