At CaMo Group, IBT, Karlsruhe Institute of Technology (KIT), I built Python pipelines for a cardiac ECG simulation study, including finite-element solves, ECG synthesis, and localization of arrhythmia origins. I ran simulations as SLURM array jobs on an HPC cluster and documented and presented the results.
I built DischargeQueue, a 30-day readmission prediction project using 97,215 US hospital discharges from an EHR database. I took it from cohort definition and feature engineering through patient-level validation and calibration, then served it through a FastAPI service with input validation and drift monitoring.
On ECG-LeadCheck, I trained a model on 21,799 clinical 12-lead ECGs to detect swapped electrodes and measured the effect of screening on false diagnostic calls. I also developed CoAtNet models for knee MRI abnormality detection in the RSNA 2026 competition.
At PsychPlus, I handled patient intake and records in eClinicalWorks and processed electronic prior authorizations through CoverMyMeds, following HIPAA procedures. My Biomedical Engineering background and experience with clinical data inform my work on healthcare predictive modeling.
