James Weatherhead

Galveston, TX · Advisor: Peter McCaffrey, MD, MS, FCAP (Chief AI Officer, UTMB)

• PhD candidate; on track to defend dissertation by December 2026, then returning to MS3 clinical rotations.
• USMLE Step 1 (Passed) · PhD Qualifying Examination (Passed) · Advanced to PhD Candidacy.

San Antonio, TX · NCAA Tennis Student-Athlete

Any query containing protected health information (PHI) must be reliably de-identified before it leaves a HIPAA-secure environment. Insufficient de-identification drives unsanctioned “shadow AI” use. The dissertation improves this boundary through three contributions:

• ASQ-PHI — an adversarial synthetic-data benchmark for clinical PHI de-identification across rule-based, cloud, vector, and on-device (Ollama) approaches (peer-reviewed; AMIA 2026 Amplify poster).
• K-anonymity decay analysis quantifying identifiability erosion across multi-turn clinical LLM conversations (peer-reviewed).
• Local PHI Scrubber — a Chrome extension for on-device, human-in-the-loop PHI redaction with 3–4B-parameter models, enabling zero data transmission (AMIA 2026 Amplify oral, Session CI25).

• Selected for a UTMB × Microsoft Research collaboration on NLP-based imaging cohort discovery.
• Built semantic search over medical image embeddings and radiology reports on Azure; multimodal pipeline across DICOM, HL7, and unstructured reports.

UTMB John Sealy School of Medicine, Galveston, TX

• Co-developed AutoTox, an AI toxicology-interpretation system in UTMB clinical pathology (Retool case study, 2024).
• Develop HIPAA-compliant semantic search and AI-agent infrastructure for clinical content discovery.
• First author on two peer-reviewed 2026 papers; co-author on a third.

San Francisco, CA

• Built biological-age models from wearable-sensor and biomarker data; designed LangGraph agents for personalized interventions.
• Integrated Junction wearable data; shipped Python to iOS/Android; traced agents with LangSmith.

UTMB, Galveston, TX · CHIRP digital health platform; CDC Social Vulnerability Index integration.

UTMB, Galveston, TX · Lipid biomarkers in neuroinflammation (murine burn-injury models); R and SPSS.

St. Mary’s University, San Antonio, TX · Anti-cancer drug testing on pediatric sarcoma cell lines; IC50 analysis.

• One of 50 winners from 4,096 global submissions (UTMB affiliation). Built VibeRad, an AI-guided DICOM viewer for medical education.
• Cited by UTMB President & CEO Jochen Reiser, MD, PhD in his June 2026 communication on UTMB’s enterprise AI initiative.

• UTMB News — “UTMB AI Innovators Win International Hackathon with Radiology Viewer and Teaching Tool” (Jun 2026).
• Microsoft Healthcare & Life Sciences Blog — “Using Natural Language to Build Healthcare Imaging Cohorts for Research” (Dec 2025).
• Retool Case Study — AutoTox reduced toxicology sign-out time 66% and increased pathology throughput tenfold (Sep 2024).

• Guest Lecturer, “Agentic Coding with Claude Code” (NTD 6088, UTMB Graduate School, Spring 2026).
• Co-President & Co-Founder, AI in Medicine Student Organization, UTMB (2022–2024).
• TMA Leadership Committee Co-Chair; AMA Committee on Health Information Technology (2023–2024).
• Member, OpenAI Forum (2026–). Kaggle competitor: NVIDIA Nemotron reasoning challenge, OpenAI agent-security red-teaming.

• Languages: Python, SQL, R.
• AI/ML: LangGraph, LangChain, LangSmith, multi-agent orchestration, vector databases, embedding models, on-device LLMs (Ollama). 4+ years applied AI/NLP; 3+ years PubMed NCBI APIs.
• Certifications: Emerald Cloud Lab Level 1 (2024); seven LangChain Academy certificates (2025–2026).