Ziad Elbagoury

Key Skills

Software

Labelbox

Scale AI

SuperAnnotate

Appen

Remotasks

Surge AI

Top Subject Matter

Computer Vision (Autonomous Driving, Retail, Drone, Medical Imaging)

Action Recognition

Surveillance Domain Expertise

Top Data Types

Image

Video

Document

Text

Audio

Top Task Types

Bounding Box

Action Recognition

Entity (NER) Classification

Classification

Transcription

RLHF

Freelancer Overview

Senior AI Data Annotator & Quality Lead—Computer Vision. Brings 14+ years of professional experience across legal operations, contract review, compliance, and structured analysis. Core strengths include Labelbox, Scale AI, and SuperAnnotate. Education includes Master of Business Administration, Canadian International College (2012) and Bachelor of Arts, Ain Shams University (2009). AI-training focus includes data types such as Image, Video, and Document and labeling workflows including Bounding Box, Action Recognition, and Entity (NER) Classification.

Labeling Experience

NLP Data Annotator—Arabic/English NLP Tasks

SuperannotateTextClassification

I labeled NLP datasets for tasks such as text classification, sentiment analysis, coreference resolution, relation extraction, and question-answer validation. Work included machine-translation quality estimation and grammatical error correction for both Arabic and English. I also reviewed and adjudicated disputed labels through consensus workflows. • Labeled and QA'd large-scale multilingual datasets. • Supported training of sentiment, translation, and relation extraction models. • Reviewed annotations in Super Annotate's consensus workflow. • Improved data quality for downstream NLP applications.

I labeled NLP datasets for tasks such as text classification, sentiment analysis, coreference resolution, relation extraction, and question-answer validation. Work included machine-translation quality estimation and grammatical error correction for both Arabic and English. I also reviewed and adjudicated disputed labels through consensus workflows. • Labeled and QA'd large-scale multilingual datasets. • Supported training of sentiment, translation, and relation extraction models. • Reviewed annotations in Super Annotate's consensus workflow. • Improved data quality for downstream NLP applications.

2022 - Present

RLHF Data Annotator & Evaluator—LLM/AI Output

Surge AITextRLHF

On RLHF and LLM evaluation tasks, I performed pairwise and ranked-choice comparisons of model responses for helpfulness, factual accuracy, tone, and safety. I wrote detailed preference rationales and calibration notes and maintained top-tier quality scores. SBS evaluations were completed efficiently while ensuring thorough and objective assessments. • Compared and ranked LLM-generated outputs for quality and safety metrics. • Authored explanations for preference decisions and calibration notes. • Conducted large-scale side-by-side evaluations on Surge AI and Invisible Technologies. • Consistently rated as a top performer for accuracy and throughput.

On RLHF and LLM evaluation tasks, I performed pairwise and ranked-choice comparisons of model responses for helpfulness, factual accuracy, tone, and safety. I wrote detailed preference rationales and calibration notes and maintained top-tier quality scores. SBS evaluations were completed efficiently while ensuring thorough and objective assessments. • Compared and ranked LLM-generated outputs for quality and safety metrics. • Authored explanations for preference decisions and calibration notes. • Conducted large-scale side-by-side evaluations on Surge AI and Invisible Technologies. • Consistently rated as a top performer for accuracy and throughput.

2022 - Present

AI Safety Content Moderator—Arabic & English

Surge AIText

I conducted AI safety and trust-and-safety evaluations, reviewing LLM and user-generated content for harm categories, policy compliance, and severity ranking. Structured incident reports and rationale documents were submitted to trust-and-safety teams. The work supported model alignment and policy updates for safer generative AI deployments. • Evaluated content for violence, self-harm, hate speech, misinformation, and PII. • Assigned severity and violation categories for trust-and-safety platforms. • Submitted rationales to contribute to policy-update cycles. • Improved alignment of large language models with responsible AI standards.

I conducted AI safety and trust-and-safety evaluations, reviewing LLM and user-generated content for harm categories, policy compliance, and severity ranking. Structured incident reports and rationale documents were submitted to trust-and-safety teams. The work supported model alignment and policy updates for safer generative AI deployments. • Evaluated content for violence, self-harm, hate speech, misinformation, and PII. • Assigned severity and violation categories for trust-and-safety platforms. • Submitted rationales to contribute to policy-update cycles. • Improved alignment of large language models with responsible AI standards.

2022 - Present

AI Transcription Annotator—Arabic & English Transcripts

RemotasksAudioTranscription

In this role, I annotated Arabic and English transcripts for speaker diarization, intent, sentiment, topic segmentation, and dialogue-act classification. I reviewed and corrected ASR outputs for dialectal Arabic and MSA, supporting conversational AI and voice-assistant training datasets. Work ensured dialect accuracy and pragmatic consistency in MENA-region deployments. • Applied sentiment and intent labels to call center and meeting transcripts. • Reviewed and corrected ASR outputs for dialectal and code-switched speech. • Segmented topics and labeled speaker turns in audio data. • Contributed to robust MENA-region conversational AI corpora.

In this role, I annotated Arabic and English transcripts for speaker diarization, intent, sentiment, topic segmentation, and dialogue-act classification. I reviewed and corrected ASR outputs for dialectal Arabic and MSA, supporting conversational AI and voice-assistant training datasets. Work ensured dialect accuracy and pragmatic consistency in MENA-region deployments. • Applied sentiment and intent labels to call center and meeting transcripts. • Reviewed and corrected ASR outputs for dialectal and code-switched speech. • Segmented topics and labeled speaker turns in audio data. • Contributed to robust MENA-region conversational AI corpora.

2022 - Present

AI Data Annotator—Email Intent and Metadata Labeling

AppenTextClassification

I conducted email and inbox annotation, extracting and labeling intent, urgency, named entities, and category information from raw email threads. The work supported AI training datasets for enterprise email-triage, auto-reply, and smart-routing systems. I flagged PII and annotated metadata fields per strict compliance standards. • Annotated Arabic and English emails for intent and action tags. • Supported training of smart email triage and routing systems. • Labeled email metadata for privacy and compliance. • Applied guidelines for consistent, high-accuracy outputs.

I conducted email and inbox annotation, extracting and labeling intent, urgency, named entities, and category information from raw email threads. The work supported AI training datasets for enterprise email-triage, auto-reply, and smart-routing systems. I flagged PII and annotated metadata fields per strict compliance standards. • Annotated Arabic and English emails for intent and action tags. • Supported training of smart email triage and routing systems. • Labeled email metadata for privacy and compliance. • Applied guidelines for consistent, high-accuracy outputs.

2022 - Present

Show more

Education

Work History