Working on applied machine learning for pediatric cardiac care — building a model to automatically detect Blalock–Taussig-Thomas (BTT) shunt thrombosis, alongside Dr. Daniel Howsmon and PhD candidate Zahra Vaez.
- Turning messy clinical signals into usable data: cleaning and preprocessing high-frequency arterial blood pressure and oxygen-saturation waveforms in Julia
- Building heatmap visualizations in Julia that reveal how frequency and amplitude shift over time, helping surface the patterns the model relies on to flag clinical events
- Turning messy clinical signals into trustworthy data — the foundation everything downstream depends on.