Shihab Al Balushi

Key Skills

Software

Appen

Clickworker

CrowdFlower

Data Annotation Tech

Google Cloud Vertex AI

iMerit

Labelbox

Lionbridge

OneForma

Sama

Scale AI

Surge AI

Toloka

Top Subject Matter

No subject matter listed

Top Data Types

Geospatial Tiled Imagery

Image

Medical Dicom

Top Task Types

Classification

Data Collection

Diagnosis

Object Detection

Text Generation

Freelancer Overview

As a multilingual NLP and Computer Vision specialist, I bring valuable experience in data annotation and AI model evaluation across multiple languages. With expertise in object detection, image labeling, and bounding box creation, I have successfully contributed to various projects at Appen, Qullit AI, and Toloka. My proficiency in Arabic, Persian, English, and German allows me to work effectively on cross-lingual projects, particularly in LLM output evaluation and prompt engineering. My background in translation combined with technical skills in annotation tools like Toloka and Appen enables me to deliver high-quality data labeling with attention to linguistic nuance and technical accuracy

Labeling Experience

Multilingual NLP Evaluation and Text Classification

Scale AITextClassificationText Generation

I contributed to a comprehensive multilingual NLP project using Scale AI's advanced annotation platform. The work involved evaluating and classifying text data across Arabic, Persian, and English languages, focusing on output comparison, ranking, and semantic accuracy review. I processed approximately 1,000 text samples, maintaining a 95% agreement rate with gold standard evaluations. The project required deep understanding of linguistic nuances across multiple languages and the ability to identify subtle semantic differences. My native fluency in Arabic and Persian was instrumental in providing high-quality annotations for these languages

I contributed to a comprehensive multilingual NLP project using Scale AI's advanced annotation platform. The work involved evaluating and classifying text data across Arabic, Persian, and English languages, focusing on output comparison, ranking, and semantic accuracy review. I processed approximately 1,000 text samples, maintaining a 95% agreement rate with gold standard evaluations. The project required deep understanding of linguistic nuances across multiple languages and the ability to identify subtle semantic differences. My native fluency in Arabic and Persian was instrumental in providing high-quality annotations for these languages

2023 - 2024

Multilingual LLM Output Evaluation and Ranking Project

AppenTextText Generation

I contributed to a comprehensive LLM evaluation project that involved analyzing and ranking outputs from various language models in multiple languages including Arabic, Persian, and English. The project required detailed assessment of semantic accuracy, contextual relevance, and linguistic nuance across approximately 500 prompt-response pairs. I performed comparative analysis between different model outputs, identified hallucinations, and evaluated response quality based on established rubrics. Quality measures included maintaining 95% agreement with gold standard evaluations and participating in regular calibration sessions to ensure consistency across the evaluation team

I contributed to a comprehensive LLM evaluation project that involved analyzing and ranking outputs from various language models in multiple languages including Arabic, Persian, and English. The project required detailed assessment of semantic accuracy, contextual relevance, and linguistic nuance across approximately 500 prompt-response pairs. I performed comparative analysis between different model outputs, identified hallucinations, and evaluated response quality based on established rubrics. Quality measures included maintaining 95% agreement with gold standard evaluations and participating in regular calibration sessions to ensure consistency across the evaluation team

2023 - 2023

Education

Work History