Toxicology education, clinical AI research, and tools for documentation and training.
PHA 6242 is a 3-credit course covering machine learning, deep learning, natural language processing, and the evaluation and validation of large language models in clinical toxicology. The 14-module curriculum connects AI methods with applications in the specialty.[1]
View the official UF course pageBylined educational case discussions on MedicalToxic, connecting clinical presentation with toxicologic reasoning.[2]
Browse case studiesA fictional interview that uses a conversational format to explain a toxic agent. An example of public-facing science communication.[3]
Read the explainerI discussed childhood lead exposure and blood testing in a Persian-language report published in 2019.[4]
View the Persian reportA Medical Toxicology podcast episode shared through my professional profile, discussing potential applications of AI in the field.[5]
View the episode announcementSelected coauthored conference papers.
Coauthored conference paper with Jafar Abdollahi · 2024 International Symposium on Telecommunications · IEEE.[6]
Coauthored conference paper with Jafar Abdollahi · QICAR 2024 · IEEE.[6]
I am CEO of MedicalToxic, a platform for clinicians and public audiences. I contribute toxicology articles, case studies and reference resources.[7]
Role: CEO and bylined contributor.
Read my articlesThe MedicalToxic team’s platform is designed to turn clinical case information into structured facts, reference-guided analysis and editable notes. It brings documentation and toxicology tools into a single case workflow.[8]
Role: MedicalToxic leadership.
Clinical effectiveness and patient-outcome benefit have not been established in the evidence reviewed for this site.
Explore ToxLumenA training module for poison-information specialists, new hires and exam candidates, with simulated consultations, practice questions and feedback.[8]
Role: A MedicalToxic team product.
Educational effectiveness has not been independently established in the evidence reviewed for this site.
View training informationCollaborative studies of machine-learning and neural-network models using National Poison Data System records. They examine whether recorded clinical features can distinguish a defined set of poisoning exposures.[9]
Role: First author on the two featured studies.
Retrospective research in selected exposure categories; prospective clinical performance remains to be tested.
Explore the studiesPossible subjects include poisoning patterns, interpretation of clinical AI research, poison-information practice and communication of toxicology evidence. Share your audience, format and learning objectives.
Start a conversationFor research collaborations, speaking and teaching invitations, editorial inquiries, and scientific consulting.