Professional Experience — AI, Localization & Language Technology
AI Data Annotator & Search Relevance Evaluator
- Contributing to the improvement of machine learning models and search algorithms through multimodal data evaluation and creation
- Analyzed and annotated search queries to refine search algorithms, ensuring strict adherence to Marketplace Relevance Project (MRP) guidelines
- Evaluated user intent by assessing Product Types (PT) and Attributes to ensure accurate product matching (Exact, Substitute, Irrelevant)
- Verified product availability and stock status across global marketplaces to maintain dataset integrity
- Contributed to Automatic Speech Recognition (ASR) datasets by recording high-fidelity voice commands in Modern Standard Arabic (MSA)
- Adhered to strict acoustic environments and pronunciation guidelines to train voice assistant models on specific keywords and wake-words
- Generated high-quality image datasets for computer vision training, including full-scene composition and isolated object photography
- Managed detailed metadata and strict file naming conventions to ensure accurate data classification for generative AI and compositing tasks
- Wrote precise editing instructions and captions to describe object-to-background relationships for model training
Media Content & Metadata Specialist
- Write detailed and structured metadata for edited videos, including final goal articulation, video descriptions, cut rationales, and audio choices
- Segment raw video footage and author rationale-rich descriptions for each segment to support clear storytelling and narrative flow
- Evaluate video content and metadata for coherence, quality, and alignment with intended visual communication
- Write detailed, descriptive captions and generate extensive, SEO-style keyword sets for media to enhance searchability and content discovery
- Assess the production quality of video and photo assets based on established criteria, including composition, lighting, focus, and adherence to brand aesthetics
- Perform content moderation by approving relevant media and rejecting unrelated or non-compliant content to maintain dataset quality and focus
- Evaluate and revise content to maintain quality, consistency, and alignment with platform standards
- Review contributor submissions, refining titles, captions, and keyword sets to ensure accuracy and relevance
- Moderate media by approving compliant content and filtering out unsuitable material
- Rank and evaluate AI-generated images by writing detailed rationales to identify the highest-quality outputs
- Collaborate with a distributed remote team, following structured workflows and communication practices
Transcriber – Arabic (Maghrebi)
- Transcribe and annotate Arabic (Maghrebi) audio content following proprietary client tool guidelines.
- Apply detailed labeling for speaker identification, PII redaction, and sound events (e.g., music, laughter, background noise)
- Follow project-specific formatting and annotation conventions to ensure dataset consistency and linguistic accuracy
- Use the Prosessor Track Sheet to manage assignments, update task status, and coordinate feedback with reviewers
- Collaborate with an international transcription team through Discord, ensuring workflow clarity and quality alignment
Linguistic Editor & Translator
- Audio Forensic Translation: Transcribed and translated sensitive audio evidence (Tunisian Arabic into English) for legal and domestic proceedings. Maintained strict verbatim accuracy, preserving emotional tone and colloquial nuances to ensure evidentiary integrity
- Socio-Political Localization: Edited and localized high-stakes content related to democracy, corruption, and governance (English into Tunisian Arabic). Successfully adapted source material from Modern Standard Arabic (MSA) into a "Formal Tunisian" register to ensure institutional relevance while maintaining local authenticity
- Dialect Consultation: Provided expert consultation on dialectal distinctions between Maghrebi varieties to support client data classification and linguistic accuracy
- Quality Assurance: Collaborated with international project managers to refine formatting and segmentation, ensuring deliverables met strict client specifications and time-sensitive deadlines
Technical Content Writer & Web Operations Assistant
- Create technical blog content and tutorials on LLM fine-tuning, NLP, and AI topics for a WordPress-based platform Author page
- Performed website audits and diagnostics to improve UX, SEO, and platform performance. View sample audit report (non-confidential)
- Collaborated with client to plan content strategy, including potential video tutorials and community engagement efforts.
- Acted as a liaison and task coordinator for external media and marketing support to ensure project delivery.
Linguistic Editor
- Conduct detailed linguistic evaluations to identify and correct translation errors in AI-generated Arabic-English content
- Assess and refine Arabic translations for clarity, accuracy, and natural fluency based on DeepL's quality standards
- Provide detailed explanations of translation issues and propose improved versions to enhance AI-driven language models
- Work on diverse language-centric projects, including categorization, labeling, transcreation, and quality assessments
- Collaborate with DeepL's Language Data team to ensure high-quality translations and AI model improvements
Linguistic AI Evaluator - Arabic Maghrebi QA | Meta AI
- Evaluate multilingual prompts and AI-generated responses in Arabic to refine advanced Large Language Models, including projects like Llama 4
- Rank outputs based on quality metrics such as coherence, fluency, and cultural relevance to enhance AI-driven conversational tools
- Provide detailed feedback to improve linguistic and contextual accuracy, ensuring alignment with diverse cultural and linguistic norms
Localization Vendor Coordinator
- Conducted real-time, scenario-based simulations with AI voicebots to evaluate conversational quality, specifically assessing turn-taking dynamics, latency, interruption handling, and error recovery
- Rated the naturalness and clarity of AI speech synthesis, providing structured feedback on pronunciation, pacing, and emotional tone to ensure voice interactions meet human-quality standards
- Evaluated and post-edited AI-generated Arabic and English content, ensuring natural fluency, contextual relevance, and cultural precision (e.g., idiomatic accuracy, tone consistency, and cultural nuance)
- Delivered and reviewed Arabic transcriptions and managed annotation and labeling tasks supporting AI model training and improvement (e.g., ASR datasets, NER tagging, and intent classification)
- Conducted audio recording sessions in Arabic and French using standardized scripts to enrich multilingual speech datasets for voice recognition and NLP development (e.g., pronunciation diversity, phonetic coverage)
- Performed auditing and evaluation of recorded audios from speakers with diverse Arabic phonetic backgrounds, assessing adherence to Modern Standard Arabic (MSA) pronunciation, grammatical accuracy, authenticity (human vs. AI), and overall data quality
- Contributed to research initiatives involving multimodal dataset creation and quality assurance for large language model (LLM) evaluation (e.g., text, audio, and video synchronization)
- Designed and reviewed audio-visual question–answer tasks to assess model reasoning, contextual understanding, and multimodal comprehension (e.g., audio-visual grounding, temporal reasoning, and scene inference)
- Documented model behavior patterns and identified areas of misinterpretation or deviation to strengthen model performance and reliability (e.g., failed reasoning cases, false correlations, or missed cues)
- Supported AI research projects evaluating the quality and accuracy of image-to-video model outputs, ensuring generated content aligns with written prompts and meets defined quality standards (e.g., motion coherence, prompt alignment, and realism assessment)
- Assisted in improving multimodal AI systems by providing structured feedback on generated media outputs and user interaction results (e.g., video response evaluation, cross-modal error logging, and benchmark refinement)
Language Data and Quality Reviewer
- Conduct comprehensive quality and safety evaluations for Generative AI and Large Language Models
- Apply complex policy guidelines to rate AI-generated responses for accuracy, truthfulness, and adherence to user intent
- Perform critical safety analysis to identify, classify, and mitigate risks from harmful content like hate speech and dangerous information
- Provide structured feedback and data annotations that directly contribute to the training and improvement of advanced AI systems
Freelance AI Data Annotator
- Conduct data labeling, image annotation, and text transcription for AI training
- Provide accurate and high-quality annotations and labels to support machine learning algorithms and AI training
- Consistently meet task requirements and quality standards set by clients and Toloka
- Collaborate with a diverse community of contributors worldwide to complete tasks efficiently
- Acquire valuable skills in data annotation, image recognition, and natural language processing
- Contribute to real-world projects across various industries, including technology, healthcare, and finance
Crowdsourced Data Contributor
- Engage in crowdsourced data collection on diverse topics using the Premise mobile app
- Provide accurate and timely information to support decision-making for governments, businesses, and organizations
- Collaborate with a global network of contributors to gather real-time insights
- Contribute to various projects including market research, infrastructure monitoring, and economic analysis
Subtitle Editor
- Translate and subtitle TED Talks into my native language, ensuring global accessibility
- Prioritize quality and accuracy while preserving the speaker's message
- Collaborate with other translators, valuing feedback for improvement
- Engage in continuous learning across diverse topics
- Contribute to the global exchange of ideas and empower non-English speakers
- Participate in community discussions and events
- Receive recognition for my contributions and commitment
KYC Validator (Know your customer)
- Review submitted identification documents for authenticity
- Cross-reference information against official records
- Maintain a high standard of accuracy and confidentiality
- Contribute to the security and reliability of the Pi Network while gaining valuable experience in regulatory compliance and identity verification processes
Freelance Editor (Arabic to English & English to Arabic)
- Translate and post-edit 1.6M+ words across 8,000+ Arabic↔English tasks
- Review and edit translations for accuracy, clarity, and style in both Arabic to English and English to Arabic
- Ensure translations maintain quality standards and linguistic integrity
- Ensure linguistic integrity and cultural appropriateness in AI-assisted translations
- Provide feedback to improve translator work and translation quality consistency
- Conduct quality checks for cultural appropriateness and accuracy to maintain high translation standards
- Collaborate with teams to deliver high-quality translations on time