I am a board-certified neurologist and physician engineer dedicated to designing, prototyping, and evaluating trustworthy clinical AI systems. Drawing from my practice in vascular neurology and neurocritical care, I focus on mitigating hallucination risks and building decision-support tools that safely augment clinical reasoning.

Rather than treating AI as an isolated technology, I look at clinical deployment through the lens of safety and automation frameworks—ensuring that LLMs and agentic workflows are deeply integrated into actual clinical workflows.

Every project in this portfolio originates from real clinical workflow friction encountered in practice. Together, these prototypes explore how AI can improve clinician usability, decision support, and healthcare workflow—from individual stroke tools to broader reimaginings of the electronic health record.

• NIHSS Mobile Assistant: A mobile-first clinical decision support prototype featuring conversational AI parsing for the NIH Stroke Scale.

• Stroke Time Tracker: A lightweight, front-end workflow tool for telestroke physicians to log and track real-time clinical milestones across multiple active cases simultaneously.

• Telestroke Dashboard: An LLM-powered quality metric extractor that pulls structured stroke metrics from unstructured text notes, with verifiable source-text tracing.

• Telestroke IVT Assistant: A clinical decision support prototype that extracts IV thrombolytic eligibility criteria from clinical notes, utilizing character-offset mapping for verifiable source-text tracing.

• Clinical Reader Mode: A concept prototype exploring an AI-powered readability layer for electronic health records that transforms cluttered clinical documentation into a standardized, clinician-centered reading experience through AI-assisted formatting, contextual linking, and intelligent information presentation.

• AI Stroke Copilot Roadmap: Modular architecture and prototypes looking toward a fully integrated command layer for acute stroke codes.

Portfolio Philosophy: These projects are modular explorations of clinical AI capabilities that could eventually be orchestrated into a comprehensive, physician-centered workflow and decision-support platform. Each prototype starts from a real clinical pain point and emphasizes safety, auditability, and practical integration into high-stakes clinical care.