AI systems architect focused on regulated enterprise workflows, insurance AI, Arabic-English evaluation, and applied machine learning.

Most teams can access the same models. The harder work is deciding what can be trusted, what must be verified, and how AI becomes an auditable system inside a real institution.

Website · Essays · Google Scholar · LinkedIn

• Evaluation design for RAG, agents, and high-accountability LLM systems.
• Insurance AI across underwriting, claims, fraud indicators, and compliance workflows.
• Arabic-English retrieval, semantic evaluation, and context engineering.
• Agentic architectures with governed tool use, audit trails, and human review.
• Computer vision and applied ML systems that move from research to operations.

• Evidence before automation.
• Human review as a designed control, not a vague fallback.
• Arabic fluency plus evidence, meaning, and decision-behavior evaluation.
• Replayable agent sessions: calls, arguments, state changes, permissions, and outputs.
• Systems that reduce institutional uncertainty rather than hiding it behind fluent interfaces.

Work on underwriting, claims, fraud indicators, audit trails, and human review. In this domain, the technical challenge is inseparable from policy interpretation, authority, exception handling, and operational adoption.

Architecture for systems where retrieval, tool calls, memory, policy constraints, model evaluation, and auditability are treated as one control surface.

Evaluation for systems that must handle Arabic and English documents, translated policy language, local terminology, dialectal intent, and evidence hierarchy.

First-author research on ensemble optimization and histology image classification, with emphasis on reproducible experimentation and model behavior under constraint.

• Histology image classification: research code related to breast-cancer histology image classification.
• Auto report automation: Python and LaTeX workflow for generating structured PDF reports from image-heavy inputs.
• Research-focused Emacs configuration: writing and research environment for long-form technical work.

• PhD in Electrical Engineering, University of Toledo.
• Four first-author peer-reviewed AI publications.
• Research focus: ensemble optimization, computer vision, histological image classification, and reproducible ML experimentation.

Selected publications are listed on Google Scholar.

From Buzzwords to Operational Change in Insurance

A keynote and executive-panel reflection from the Global Insurance Innovation Forum Saudi Arabia 2025 on why insurance transformation depends on data quality, phased integration, change management, and operational AI controls.