Senior at UC Irvine studying data science with a minor in management. I build software that puts machine learning into production — from RAG pipelines and AI agents to APIs and analytics infrastructure. Most recently at Chipotle, where I cut a four-hour reporting workflow down to twenty minutes.
A layered C# / .NET 10 system handling real insurance math — copay → deductible → coinsurance → out-of-pocket maximum — with 6 business-rule validations, 10 REST endpoints, and 18 unit tests. Containerized, CI/CD'd, and deployed live.
End-to-end RAG pipeline ingesting real 10-K filings from SEC EDGAR. 220k+ characters embedded via Gemini, stored in ChromaDB. Top-5 semantic retrieval with cited answers — eliminating hallucinations by grounding responses in source documents.
A 3-model ensemble — Isolation Forest, LSTM Autoencoder, DBSCAN — over 2.5M+ NSE/BSE F&O records. Without labels, the system independently flagged Sept 20, 2019 as the most anomalous NIFTY day, correlating with the Saudi Aramco oil attack four days prior.
Four-model credit scoring system on 391k LendingClub loans. SHAP TreeExplainer surfaces per-loan reasoning; CFPB 4/5ths rule audit detected disparate impact (ratio: 0.68) against low-income applicants. Findings live in an interactive Dash dashboard.
Full-stack developer tool syncing GitHub PRs and issues in real time into a Kanban board UI. Built with GitHub OAuth for secure access, REST endpoints for internal services, and CI/CD via GitHub Actions — deployed live and actively used for project visibility.
Analyzed 1.5M+ rows of retail transaction data to surface churn and growth signals hidden in raw sales history. Built executive-facing dashboards in Power BI and Tableau, translating a hidden churn spike and a major revenue-growth opportunity into decision-ready reporting.