Medical Doctor · AI Operations Specialist · Theological Essayist
A hybrid practitioner combining bedside clinical training, AI evaluation operations, and healthcare education — graduate of Zagazig Faculty of Medicine and published systematic reviewer turning clinical expertise into defensible data that powers safer medical AI. Also writing long-form essays on theology, philosophy & depth psychology.
Where clinical depth meets precision-engineered AI evaluation.
I'm a graduate of Zagazig Faculty of Medicine, a published systematic reviewer, and an AI training data specialist working at the intersection of clinical medicine, payer policy, and large language models. My evaluations don't just check whether an answer is "correct" — they assess whether it is legally defensible, clinically safe, and ready for real-world deployment.
I've evaluated medical prior-authorization AI responses involving oncology pharmacotherapy, molecular mutation analysis, and step-therapy compliance — led QA on Arabic LLM training data, audio transcription batches, and a 100-item SEC financial Q&A annotation pipeline I designed end-to-end — and serve as International Affairs Strategic Lead at GAIA, an Egyptian agri-food startup.
Graduate of Zagazig Faculty of Medicine with clinical rotations across internal medicine, surgery, pediatrics, and psychiatry. I evaluate AI outputs the way a Medical Director of Pharmacy Operations would — cross-referencing payer policies, NGS reports, and specialist notes in real time.
Clinical & Medical Research
Hands-on clinical research across behavioral medicine, personality disorders, developmental adversity, and psychometric methodology. Comfortable moving between bedside reasoning, literature synthesis, and quantitative analysis.
Published Systematic Reviewer
Co-author on a PRISMA systematic review & meta-analysis on childhood emotional adversity and vulnerable narcissism — 51 studies synthesized, pooled r = 0.42, I² = 86%, Egger's test p = 0.28. Under review at Personality and Individual Differences (Elsevier, PAID-D-26-00409R1).
AI Training & Annotation
Multi-dimensional rubrics for accuracy, reasoning, completeness, and safety. Arabic and English LLM evaluation across formal, neutral, and conversational registers.
QA Pipelines That Hold Up
Style guides, error taxonomies, rater screening tests, and rolling QA cycles — designed so every dataset entry is independently verifiable, machine-parseable, and defensible on appeal.
International Affairs — GAIA
International Affairs Strategic Lead at GAIA, an Egyptian agri-food startup. Built the full international communications package: intelligence brief, investor narrative one-pager, WFF Startup Innovation Awards 2026 application, and an Arabic-language patent protection proposal.
Theological & Philosophical Essayist
Independent long-form writing at the intersection of Patristic theology, existential philosophy, and depth psychology. Themes include the divided self, narcissism as flight from being, the structure of love, and a liturgical-psychological reading of Holy Week. Published at Divine Philosophy.
Services
What I deliver
Specialized services for AI labs, healthcare-AI startups, and annotation vendors that need clinically defensible training data and evaluations.
Medical LLM Evaluation
Physician-level review of LLM responses against payer policy, clinical guidelines, NGS reports, and specialist notes.
Clinical Hallucination Detection
Catch fabricated drug indications, wrong dosing, false guideline citations, and unsafe clinical reasoning before they ship.
Arabic LLM QA
MSA + Egyptian colloquial. Formal, neutral, and conversational registers — with structured correction pathways.
Medical Transcription QA
4-tier error taxonomy (Minor / Moderate / Significant / Critical) with timestamped citations and word-level accuracy scoring.
Annotation Guidelines
Style guides, edge-case libraries, validation rules, and rater onboarding workflows that hold up under audit.
Rubric Design
Multi-dimensional rubrics calibrated for inter-rater reliability — built around accuracy, reasoning, completeness, and patient safety.
Need medical AI outputs reviewed for clinical safety, hallucinations, or policy compliance?
Send me a sample task. I'll return a physician-level evaluation with a clear rubric, error taxonomy, and revision pathway.
AlignerrUpworkGoogle Sheets / ExcelNotionLabel StudioChatGPT / Claude / Gemini
Featured Work
Selected Projects
Real evaluation samples and pipeline designs from medical AI, Arabic LLM QA, audio transcription, and academic research.
Medical AI · Oncology
Medical Prior-Authorization AI Evaluation
Evaluated AI-generated prior-authorization determinations for a CML / Bosutinib case at Medical Director of Pharmacy Operations level. Cross-referenced payer policy bulletin (OHS-CML-2024-004B), NGS reports (T315I mutation, VAF 18.4%), bone marrow biopsy, and 15+ specialist notes. Identified a prescriber clinical error on Bosutinib efficacy for T315I that the weaker AI response missed entirely — and chose the response that built a legally defensible denial via verbatim policy citation and a proactive Clinical Alert.
A+++ preference call15+ source documentsVerbatim citation standard
SEC Financial Q&A Dataset — Pipeline Designer & QA Lead
Designed a complete annotation pipeline for a 100-item AI training dataset of SEC 10-K / 10-Q financial Q&A pairs. Built the 3-tier category framework (A: single-fact, B: multi-fact synthesis, C: cross-document comparison), verbatim citation standard with mandatory page numbers, 30-point rater screening rubric, rolling QA cycle, and automated validation rules — all so every entry is independently verifiable and machine-parseable from day one.
Evaluated AI-generated Arabic content across formal, neutral, and conversational registers — including MSA and Egyptian colloquial. Flagged factual errors, grammatical inconsistencies, and stylistic misalignments, producing structured written feedback with specific correction pathways for data contributors.
95%+ accuracyAll registersBilingual feedback
Arabic NLPRegister AnalysisLinguistic QA
Research · Elsevier
Systematic Review & Meta-Analysis — Elsevier
Co-authored a PRISMA-compliant systematic review on childhood emotional adversity and vulnerable narcissism, currently under review at Personality and Individual Differences (Elsevier, PAID-D-26-00409R1). Led full-text screening, dual extraction, ROBINS-E risk-of-bias assessment, random-effects meta-analysis, and manuscript drafting. Synthesized 51 studies — pooled r = 0.42, I² = 86%, Egger's test p = 0.28.
Ongoing literature reviews and research synthesis at the intersection of clinical medicine and psychology — covering personality disorders, developmental adversity, psychometric methodology, and behavioral medicine. Comfortable with PubMed search strategy, critical appraisal, and translating findings for both academic and clinical audiences.
3 active research domainsBilingual literaturePubMed / Cochrane workflow
Literature ReviewCritical AppraisalPsychometricsClinical Synthesis
Strategy · GAIA
GAIA — International Affairs Strategic Lead
Built the full international communications package for GAIA, an Egyptian agri-food startup: country/market intelligence brief, investor narrative one-pager, WFF Startup Innovation Awards 2026 application materials, and an Arabic-language patent protection proposal. Translate complex scientific and business concepts into clear narratives for international funders, partners, and stakeholders.
Authored a 6-dimension rubric for evaluating AI responses in medical utilization review: Policy Precision, Evidence Specificity, Error Detection, Completeness, Actionability, and Patient-Safety Priority — drawn from real-world PA adjudication standards and payer policy requirements.
Outside the clinic and the evaluation queue, I write long-form essays at the intersection of Patristic theology, existential philosophy, and depth psychology. The essays read Scripture, the Church Fathers, and the human person as one continuous text — asking how shame, love, narcissism, and resurrection actually work inside a real life.
Rooted in the Alexandrian and Cappadocian fathers — Clement, Origen, Gregory of Nyssa, and Maximus the Confessor — recovering the mystical-philosophical line that often gets lost behind later polemics.
Existential Register
Theology read through being, freedom, shame, and the divided self. Less doctrinal exposition, more an attempt to ask what it actually means to exist before God and before another person.
Depth-Psychological Lens
Narcissism as flight from being, the self-image formed in childhood, the mechanics of love and rejection. The clinical eye and the contemplative eye reading the same human at the same time.
Conversational Arabic Prose
Long-form essays in colloquial-leaning Arabic — written as if thinking out loud with the reader. Slow rhythm, parenthetical asides, and a refusal to flatten complexity into slogans.
Need medical AI outputs reviewed for clinical safety, hallucinations, or policy compliance? Send me a sample task — I'll return a physician-level evaluation with a clear rubric and revision pathway.
