Dark storm clouds with blue evening light

Projects

A collection of independent work exploring ideas across technology, business, markets, and design. These projects range from products I build and test to research, analysis, and written work, each driven by curiosity and an interest in turning ideas into something tangible.

Ambient Iris concept showing a dark workspace and illuminated neural presence

Local-first personal AI workspace | Native macOS assistant | v0.11.6 public snapshot

AXIOM / Iris

AXIOM / Iris is a local-first personal intelligence system built around one idea: a single conversation should be able to consult the right documents, schedule, coursework, reminders, and connected information without turning the user into the integration layer. Iris is the conversational experience; AXIOM is the system behind it, combining a Python-based local system, native macOS client, approved source collections, speech and language models, local storage, and scoped service connections.

Current prototype supports typed and voice conversation, approved-document search, passage-backed answers, source preview, calendar/task/reminder workflows, web/current-information lookup, and editable memory.
Design emphasis: source receipts, visible control, honest capability boundaries, reversible preferences, and recovery paths when speech, context, or connected services fail.
Public scope: product case study and approved visuals only; app source, prompts, credentials, private data, and personal records remain unpublished.

GitHub Project PDF Case Study

Weather-aware wardrobe system | iPad-first personal product beta

Drape

Drape is a cross-device wardrobe system that turns clothes already owned, local weather, and personal preferences into practical daily outfit recommendations. Originally designed for an iPad on a stand, it combines a calm dashboard, digital wardrobe, garment photo tools, manual outfit canvas, saved looks, and a local rule-based styling engine that keeps outfit logic explainable instead of hiding it behind a black box.

Built with a shared JavaScript interface, SwiftUI/WKWebView Apple shell, Supabase account syncing, Cloudflare hosting, and Apple Vision support on native devices.
Core workflows include clothing upload, category/layer organization, searchable wardrobe, weather-aware suggestions, layered outfits, saved looks, and editable item metadata.
Current stage: working personal beta, not an App Store release; image recognition and recommendation quality still require broader testing and user review.

Live Demo GitHub Project PDF