Things I've built.

Storefront Next is Salesforce's next-generation framework for building headless B2C Commerce storefronts. Built on React 19 and React Router 7, it combines streaming server-side rendering with server-side personalization, and ships as production-ready templates that deploy to Salesforce Managed Runtime. I lead core engineering initiatives on the framework.

• Streaming server-side rendering on React 19 and React Router 7, with file-based routing and full Commerce API (SCAPI) integration.
• Server-side personalization with real-time shopper context, kept in sync with stateless SLAS OAuth sessions.
• Hybrid Authentication, which I co-authored, lets merchants move to React incrementally with zero downtime.
• Managed Runtime environment orchestration that I designed, cutting the engineering feedback loop from 3 hours to 7 minutes.
• Six vertical templates, each scaffolded from one CLI with a single --vertical flag.

Reflections turns my photo library into a magazine-style analytics dashboard. A nightly pipeline ingests every photo and derives signals like EXIF data, color palettes, semantic embeddings, style clusters and composition, then serves them to a web app and a native iOS app. Spin the globe to find where each photo was taken, or browse by trip, year and collection.

• Python ingest pipeline on GitHub Actions that mirrors every photo to Cloudflare R2 at three sizes and writes its metadata to Neon Postgres.
• Semantic search: SigLIP 2 embeddings stored in pgvector, with a FastAPI and ONNX text encoder ranking photos by similarity.
• Style clustering with scikit-learn, color-palette extraction and saliency-based composition analysis.
• React Router 7 web app with a versioned JSON API that also powers a read-only Expo iOS app.
• Owner-only admin behind Clerk, and OpenTelemetry tracing into Axiom across the web app and pipeline.

The Progressive Web App Kit is Salesforce's framework for building storefronts with React on the Commerce API. It gives front-end teams a flexible, modern alternative to traditional storefront architecture.

• Project templates, including the Retail React App: a customizable storefront covering core commerce flows from home page to checkout.
• A rendering system that runs on both server and client, including hydration.
• A routing system that injects Commerce API data into components.
• Utilities and scripts that automate routine development tasks.

I/O Sense is a cloud IoT platform built to accelerate digital transformation for enterprises. It's configured for energy management, water management, temperature monitoring and more, and serves as an Industrial IoT (IIoT) platform for digitizing processes and assets.

• Gives operations teams visibility into processes and assets in real time.
• Automates operations and adds process analytics for data-driven decisions.
• Industrial IoT capabilities for digitizing and predicting the behavior of processes and assets.

A secure messaging system modeled on Slack, built as a graduate software engineering project. Group messaging, secure conversations, file exchange, threaded replies and accessible front-end clients were at its core.

• Group and direct messaging with thread-based conversations.
• Secure messaging and file exchange.
• Java and Spring Boot back end, tested with JUnit.