Research tools, statistical applications, and infrastructure built across clinical and academic contexts.
A collection of interactive Shiny dashboards supporting ongoing IRB-approved studies in pediatric trauma, neurocritical care, and vascular surgery. Each dashboard is built for study collaborators and supports real-time data exploration, statistical model outputs, and injury phenotype classification. Underlying analyses span mixed-effects probit regression, recursive partitioning, and annotated anatomical SVG visualization.
Access restricted to authorized study collaborators
Self-hosted infrastructure designed to support research workflows involving sensitive clinical data. Key design decisions: local LLM inference via llama.cpp and a Qwen model for PHI-safe AI-assisted analysis; Claude Code on WSL2 for de-identified work; Ansible-managed baseline hardening across a heterogeneous Linux cluster.
Monitoring stack includes Grafana, Prometheus, Loki, and Uptime Kuma for unified observability across Ubuntu, Fedora, and AlmaLinux nodes with mixed Docker/Podman runtimes.
Faculty instructor at two institutions: the San Diego State University Graduate School of Public Health and the UC San Diego School of Medicine Clinical Research Enhancement through Supplemental Training (CREST) Program.
At SDSU, I teach Data Analysis with R — a graduate-level course covering data wrangling, visualization, and statistical modeling in R. At UCSD CREST, I teach Epidemiology I and Applied Epidemiology, covering study design, causal inference, and the application of epidemiologic methods to clinical research questions. I also serve on thesis and dissertation committees at both campuses.
Custom websites designed and built for clinical and academic contexts, with an emphasis on professional presentation, typographic clarity, and site performance.