A document and data portal that turns a real estate development firm's scattered Dropbox spreadsheets into an operating system.

Traditional real estate firms run on information scattered across Dropbox folders, email threads, and one-off spreadsheets — nobody outside the person who built a given Excel model can see what it says. I built a portal that ingests those scattered documents straight out of Dropbox and turns them into live dashboards, budget approval workflows, and underwriting reviews, giving the company one place to see real data across the whole lifecycle: land acquisition, due diligence, underwriting, budgeting, and ongoing asset management.

Type any /command, press space, and its stored text appears — in any app, in any text field.

A desktop tray app that expands /commands into saved text anywhere on the system: a Rust keyboard hook watches for the trigger and rewrites it inline. I built it while applying for jobs, to stop retyping my phone number and email into form after form. The interesting part isn't the snippet manager — it's intercepting and rewriting keystrokes inside someone else's application without breaking their text field.

An iOS app that logs a watering when you tap the NFC sticker on the pot — and nudges you when you forget.

An iOS app for tracking plant care: tap an NFC sticker on the pot to log a watering, and get a push notification when it's overdue. Built for myself after I kept forgetting to water my own plants — the interesting part is that the physical tap replaces the usual chore of opening an app and finding the right plant, so logging is as fast as the watering itself.

An NLP study of what Tesla's and Rivian's own language reveals about how each frames risk, performance, and outlook.

A financial NLP pipeline that extracts, cleans, and scores three text sources per company — 10-K Item 1A and Item 7 for FY2023 and FY2024, earnings call transcripts split into management discussion and Q&A, and CEO/CFO prepared remarks — then compares Tesla against Rivian on sentiment, readability, year-over-year similarity, and topic extraction.