Nourhan Makhlouf

Key Skills

Software

AWS SageMaker

Clickworker

CloudFactory

CVAT

Data Annotation Tech

Google Cloud Vertex AI

Labelbox

OneForma

Toloka

Telus

V7 Labs

Top Subject Matter

No subject matter listed

Top Data Types

Image

Text

Video

Top Task Types

Classification

Diagnosis

Evaluation Rating

Object Detection

RLHF

Freelancer Overview

I have a strong foundation in data labeling and AI training, particularly in projects that involve medical content, bilingual text processing (Arabic-English), and search relevance evaluation. With a background in pharmacy and fluency in both Arabic and English, I bring domain-specific knowledge to healthcare-related annotation tasks, such as labeling pharmaceutical texts, medical terminology, and clinical documentation. My attention to detail and scientific accuracy ensures high-quality training data for models used in healthcare AI applications. Additionally, I have contributed to natural language processing tasks including machine translation alignment, sentiment analysis, and named entity recognition, especially within Arabic-English datasets. I am also familiar with various annotation platforms and tools used for text, image, and search evaluation tasks. My ability to work accurately with specialized content and across languages makes me a strong asset in projects requiring both linguistic precision and subject-matter expertise.

Labeling Experience

Medical Image Annotation for Diagnostic AI

CVATMedical DicomBounding BoxSegmentation

Annotated a wide range of medical images, including X-rays, CT scans, and histopathology slides, to support the training of AI models for diagnosis and disease detection. Tasks included drawing bounding boxes and segmentation masks around abnormalities (e.g., tumors, lesions, fractures), and labeling image sets with corresponding diagnoses or anatomical regions. Collaborated with clinical guidelines and reference data to ensure annotation consistency. Followed strict quality assurance protocols including inter-annotator validation and consensus review processes to meet regulatory-grade data standards.

Annotated a wide range of medical images, including X-rays, CT scans, and histopathology slides, to support the training of AI models for diagnosis and disease detection. Tasks included drawing bounding boxes and segmentation masks around abnormalities (e.g., tumors, lesions, fractures), and labeling image sets with corresponding diagnoses or anatomical regions. Collaborated with clinical guidelines and reference data to ensure annotation consistency. Followed strict quality assurance protocols including inter-annotator validation and consensus review processes to meet regulatory-grade data standards.

2023 - 2024

Search Relevance Evaluation

TolokaTextEntity Ner ClassificationClassification

Worked on evaluating search relevance across multiple industries, including e-commerce, general knowledge, and online content categorization. Evaluated and categorized search results for queries spanning a wide range of topics such as fashion, electronics, automotive, entertainment, and general consumer products. Tasks included analyzing user intent, ensuring that search results were aligned with the users' specific needs, and rating search results for accuracy, relevance, and trustworthiness. Also helped assess search ranking algorithms to ensure that the most useful and high-quality content appeared first for diverse types of searches. Focused on improving user experience and the relevance of results in real-world, multi-field search scenarios.

Worked on evaluating search relevance across multiple industries, including e-commerce, general knowledge, and online content categorization. Evaluated and categorized search results for queries spanning a wide range of topics such as fashion, electronics, automotive, entertainment, and general consumer products. Tasks included analyzing user intent, ensuring that search results were aligned with the users' specific needs, and rating search results for accuracy, relevance, and trustworthiness. Also helped assess search ranking algorithms to ensure that the most useful and high-quality content appeared first for diverse types of searches. Focused on improving user experience and the relevance of results in real-world, multi-field search scenarios.

2021 - 2024

Arabic-English Audio Transcription and Voice Recording for Speech AI

TolokaAudioClassificationEmotion Recognition

Contributed to multilingual speech datasets by recording Arabic and English voice samples for use in speech recognition and voice assistant training. Performed high-quality transcription of spoken content, including conversational, instructional, and medical dialogues. Labeled transcripts with speaker ID, background noise tags, and emotional tone where applicable. Projects followed strict linguistic and acoustic guidelines to ensure phonetic diversity, dialect representation, and transcription accuracy. Delivered clean, timestamped transcriptions and validated pronunciation and clarity across dialects (including Egyptian Arabic).

Contributed to multilingual speech datasets by recording Arabic and English voice samples for use in speech recognition and voice assistant training. Performed high-quality transcription of spoken content, including conversational, instructional, and medical dialogues. Labeled transcripts with speaker ID, background noise tags, and emotional tone where applicable. Projects followed strict linguistic and acoustic guidelines to ensure phonetic diversity, dialect representation, and transcription accuracy. Delivered clean, timestamped transcriptions and validated pronunciation and clarity across dialects (including Egyptian Arabic).

2021 - 2024

Bilingual LLM Evaluation – Arabic and English

TolokaTextClassificationQuestion Answering

Performed human evaluation and alignment testing for Arabic and English prompts and responses generated by large language models. Tasks included scoring AI outputs for relevance, fluency, factual accuracy, coherence, bias, and harmful content, across both general and domain-specific topics. Participated in prompt writing and fine-tuning evaluations to improve LLM performance in Arabic. Also contributed to Red Teaming tasks aimed at identifying safety vulnerabilities and hallucinations in model responses. Maintained consistency and linguistic accuracy across both languages, adhering to quality control standards such as agreement scoring and calibration rounds.

Performed human evaluation and alignment testing for Arabic and English prompts and responses generated by large language models. Tasks included scoring AI outputs for relevance, fluency, factual accuracy, coherence, bias, and harmful content, across both general and domain-specific topics. Participated in prompt writing and fine-tuning evaluations to improve LLM performance in Arabic. Also contributed to Red Teaming tasks aimed at identifying safety vulnerabilities and hallucinations in model responses. Maintained consistency and linguistic accuracy across both languages, adhering to quality control standards such as agreement scoring and calibration rounds.

2021 - 2024

Education

Work History