Computer science at UC Berkeley. I build ML systems and do research in computer vision and medical AI — robustness, bias, augmentation, segmentation.
Computer vision for medical imaging — robustness, reliability, and the failure modes that don’t show up on the average. The full publication list lives on the Research page.
Physiology-guided augmentation pipelines that explicitly model breathing motion and anatomical deformation to improve robustness of medical segmentation.
Hybrid analytical + deep-learning approach for generating stereoscopic pairs from a single image.
Segmentation model augmented with GAN-generated ecological context to improve cross-background generalization.
Multimodal pricing engine (vision + NLP) and large-scale ingestion pipeline for resale marketplaces.
I was curious what a Kalshi specifically for Cal events would look like, so I built it. Reached 100 users on launch.
A vision-language project exploring multimodal representations and evaluation workflows.
A lightweight speed-reading tool using rapid serial visual presentation (RSVP) to improve reading efficiency while maintaining comprehension.
Email is the best way to reach me.
.png)
