ML / Data Co-op
Modeling fraud signals across consumer and commercial ACH data.
Exploratory analysis on fraud and ACH transactions, with stakeholder dashboards for fraud-probability review.
From a first software internship to AI/ML research and a GSoC mentorship — it starts at the bottom, in 2024, and climbs to the present.
Modeling fraud signals across consumer and commercial ACH data.
Exploratory analysis on fraud and ACH transactions, with stakeholder dashboards for fraud-probability review.
Physics-informed deep learning for strong gravitational lensing.
Turned a ResNet into a PINN for dark-matter classification (AUC 0.994 → 0.9953 on an H200 cluster) and built a self-supervised ViT pipeline reaching 0.999 macro AUC.
Automated reporting pipelines that cut manual data handling by 70%.
Python/SQL ETL, Power BI dashboards for lease managers (~160 hrs/yr saved), SSRS reporting, and test→prod deployments. ~$50K annual savings.
Shipped a software-asset tracking feature across 11 production sites.
Owned requirements and success criteria, wrote Pytest unit tests, and contributed 10 PRs and 14 fixes across 16 internal Django apps. ~$63K/year in savings.